National Academies Press: OpenBook

Utilization Measurement and Management of Fleet Equipment (2021)

Chapter: Part I - Research Overview

« Previous: Front Matter
Page 1
Suggested Citation:"Part I - Research Overview." National Academies of Sciences, Engineering, and Medicine. 2021. Utilization Measurement and Management of Fleet Equipment. Washington, DC: The National Academies Press. doi: 10.17226/26067.
×
Page 1
Page 2
Suggested Citation:"Part I - Research Overview." National Academies of Sciences, Engineering, and Medicine. 2021. Utilization Measurement and Management of Fleet Equipment. Washington, DC: The National Academies Press. doi: 10.17226/26067.
×
Page 2
Page 3
Suggested Citation:"Part I - Research Overview." National Academies of Sciences, Engineering, and Medicine. 2021. Utilization Measurement and Management of Fleet Equipment. Washington, DC: The National Academies Press. doi: 10.17226/26067.
×
Page 3
Page 4
Suggested Citation:"Part I - Research Overview." National Academies of Sciences, Engineering, and Medicine. 2021. Utilization Measurement and Management of Fleet Equipment. Washington, DC: The National Academies Press. doi: 10.17226/26067.
×
Page 4
Page 5
Suggested Citation:"Part I - Research Overview." National Academies of Sciences, Engineering, and Medicine. 2021. Utilization Measurement and Management of Fleet Equipment. Washington, DC: The National Academies Press. doi: 10.17226/26067.
×
Page 5
Page 6
Suggested Citation:"Part I - Research Overview." National Academies of Sciences, Engineering, and Medicine. 2021. Utilization Measurement and Management of Fleet Equipment. Washington, DC: The National Academies Press. doi: 10.17226/26067.
×
Page 6
Page 7
Suggested Citation:"Part I - Research Overview." National Academies of Sciences, Engineering, and Medicine. 2021. Utilization Measurement and Management of Fleet Equipment. Washington, DC: The National Academies Press. doi: 10.17226/26067.
×
Page 7
Page 8
Suggested Citation:"Part I - Research Overview." National Academies of Sciences, Engineering, and Medicine. 2021. Utilization Measurement and Management of Fleet Equipment. Washington, DC: The National Academies Press. doi: 10.17226/26067.
×
Page 8
Page 9
Suggested Citation:"Part I - Research Overview." National Academies of Sciences, Engineering, and Medicine. 2021. Utilization Measurement and Management of Fleet Equipment. Washington, DC: The National Academies Press. doi: 10.17226/26067.
×
Page 9
Page 10
Suggested Citation:"Part I - Research Overview." National Academies of Sciences, Engineering, and Medicine. 2021. Utilization Measurement and Management of Fleet Equipment. Washington, DC: The National Academies Press. doi: 10.17226/26067.
×
Page 10
Page 11
Suggested Citation:"Part I - Research Overview." National Academies of Sciences, Engineering, and Medicine. 2021. Utilization Measurement and Management of Fleet Equipment. Washington, DC: The National Academies Press. doi: 10.17226/26067.
×
Page 11
Page 12
Suggested Citation:"Part I - Research Overview." National Academies of Sciences, Engineering, and Medicine. 2021. Utilization Measurement and Management of Fleet Equipment. Washington, DC: The National Academies Press. doi: 10.17226/26067.
×
Page 12
Page 13
Suggested Citation:"Part I - Research Overview." National Academies of Sciences, Engineering, and Medicine. 2021. Utilization Measurement and Management of Fleet Equipment. Washington, DC: The National Academies Press. doi: 10.17226/26067.
×
Page 13
Page 14
Suggested Citation:"Part I - Research Overview." National Academies of Sciences, Engineering, and Medicine. 2021. Utilization Measurement and Management of Fleet Equipment. Washington, DC: The National Academies Press. doi: 10.17226/26067.
×
Page 14
Page 15
Suggested Citation:"Part I - Research Overview." National Academies of Sciences, Engineering, and Medicine. 2021. Utilization Measurement and Management of Fleet Equipment. Washington, DC: The National Academies Press. doi: 10.17226/26067.
×
Page 15
Page 16
Suggested Citation:"Part I - Research Overview." National Academies of Sciences, Engineering, and Medicine. 2021. Utilization Measurement and Management of Fleet Equipment. Washington, DC: The National Academies Press. doi: 10.17226/26067.
×
Page 16
Page 17
Suggested Citation:"Part I - Research Overview." National Academies of Sciences, Engineering, and Medicine. 2021. Utilization Measurement and Management of Fleet Equipment. Washington, DC: The National Academies Press. doi: 10.17226/26067.
×
Page 17
Page 18
Suggested Citation:"Part I - Research Overview." National Academies of Sciences, Engineering, and Medicine. 2021. Utilization Measurement and Management of Fleet Equipment. Washington, DC: The National Academies Press. doi: 10.17226/26067.
×
Page 18
Page 19
Suggested Citation:"Part I - Research Overview." National Academies of Sciences, Engineering, and Medicine. 2021. Utilization Measurement and Management of Fleet Equipment. Washington, DC: The National Academies Press. doi: 10.17226/26067.
×
Page 19
Page 20
Suggested Citation:"Part I - Research Overview." National Academies of Sciences, Engineering, and Medicine. 2021. Utilization Measurement and Management of Fleet Equipment. Washington, DC: The National Academies Press. doi: 10.17226/26067.
×
Page 20
Page 21
Suggested Citation:"Part I - Research Overview." National Academies of Sciences, Engineering, and Medicine. 2021. Utilization Measurement and Management of Fleet Equipment. Washington, DC: The National Academies Press. doi: 10.17226/26067.
×
Page 21
Page 22
Suggested Citation:"Part I - Research Overview." National Academies of Sciences, Engineering, and Medicine. 2021. Utilization Measurement and Management of Fleet Equipment. Washington, DC: The National Academies Press. doi: 10.17226/26067.
×
Page 22
Page 23
Suggested Citation:"Part I - Research Overview." National Academies of Sciences, Engineering, and Medicine. 2021. Utilization Measurement and Management of Fleet Equipment. Washington, DC: The National Academies Press. doi: 10.17226/26067.
×
Page 23
Page 24
Suggested Citation:"Part I - Research Overview." National Academies of Sciences, Engineering, and Medicine. 2021. Utilization Measurement and Management of Fleet Equipment. Washington, DC: The National Academies Press. doi: 10.17226/26067.
×
Page 24
Page 25
Suggested Citation:"Part I - Research Overview." National Academies of Sciences, Engineering, and Medicine. 2021. Utilization Measurement and Management of Fleet Equipment. Washington, DC: The National Academies Press. doi: 10.17226/26067.
×
Page 25
Page 26
Suggested Citation:"Part I - Research Overview." National Academies of Sciences, Engineering, and Medicine. 2021. Utilization Measurement and Management of Fleet Equipment. Washington, DC: The National Academies Press. doi: 10.17226/26067.
×
Page 26
Page 27
Suggested Citation:"Part I - Research Overview." National Academies of Sciences, Engineering, and Medicine. 2021. Utilization Measurement and Management of Fleet Equipment. Washington, DC: The National Academies Press. doi: 10.17226/26067.
×
Page 27
Page 28
Suggested Citation:"Part I - Research Overview." National Academies of Sciences, Engineering, and Medicine. 2021. Utilization Measurement and Management of Fleet Equipment. Washington, DC: The National Academies Press. doi: 10.17226/26067.
×
Page 28
Page 29
Suggested Citation:"Part I - Research Overview." National Academies of Sciences, Engineering, and Medicine. 2021. Utilization Measurement and Management of Fleet Equipment. Washington, DC: The National Academies Press. doi: 10.17226/26067.
×
Page 29
Page 30
Suggested Citation:"Part I - Research Overview." National Academies of Sciences, Engineering, and Medicine. 2021. Utilization Measurement and Management of Fleet Equipment. Washington, DC: The National Academies Press. doi: 10.17226/26067.
×
Page 30

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Research Overview P A R T I

C O N T E N T S 3 Summary 8 Chapter 1 Introduction 8 1.1 Background 8 1.2 Project Objective 8 1.3 Research Approach 8 1.4 Organization of the Report 9 Chapter 2 Research Approach 9 2.1 Introduction 9 2.2 Literature Review 9 2.3 Agency Survey 9 2.4 Data Collection 10 2.5 Data Processing 10 2.6 Fitting the Models 10 2.7 Validating the Models 11 2.8 Developing the Equipment Fleet Utilization Management Program 11 2.9 Validating the Equipment Fleet Utilization Management Program 11 2.10 Developing the UPM Software 11 2.11 Validating the UPM Software 12 Chapter 3 Research Findings 12 3.1 Introduction 12 3.2 Literature Review 15 3.3 Agency Survey 17 3.4 Data Collection 19 3.5 Utilization Estimation Models 21 3.6 Equipment Cost Estimation Models 21 3.7 Validating the Models 24 3.8 Developing the Equipment Fleet Utilization Management Program 26 3.9 Validating the Equipment Fleet Utilization Management Program 27 Chapter 4 Summary and Suggested Research 27 4.1 Summary 28 4.2 Suggested Research 29 References

3 S U M M A R Y Background Equipment fleet assets are vital to the delivery of state highway agency programs, projects, and services. Measuring, monitoring, and reporting on equipment utilization levels are necessary for the management of equipment fleet and meeting highway agencies’ business needs. Highway agencies use a variety of processes for utilization measurement and manage- ment; there is no widely accepted process for determining utilization criteria, measurement, and management of fleet equipment. There was a need to develop a guide that incorporates rational processes and appropriate electronic-based tools and provides a realistic means for fleet utilization measurement and management. Such a guide will help equipment managers and administrators in making decisions regarding fleet size and composition to meet their agency missions. NCHRP Project 13-05 was initiated to address this need. Objective The objective of this research was to develop a guide for utilization measurement and management of fleet equipment and associated processes and electronic-based tools for use by state highway agencies. Findings Literature Review and Survey The literature search revealed availability of limited information directly related to fleet equipment utilization. Relevant information was obtained from fleet management practitioners at state departments of transportation (DOTs); 32 responses were received. The literature review and survey showed that annual mileage, annual engine hours, usage over the last 12 months, and frequency of use were the most widely used utilization metrics. The factors that influenced utilization measurement and management (directly or indirectly) were equipment in-service age, cumulative utilization level, breakdown rate, downtime hours, fleet size, number of equipment units under repair, equipment quality upon acquisition, normal equipment life, expected equipment workload, supervision level needed for equipment, product types, environmental impacts, purchase cost, operating cost, maintenance cost, repair cost, and salvage value. The survey results indicated that the equipment types most commonly used by state DOTs were dump trucks, pickup trucks, automobiles, vans, sport utility vehicles, trailers, Utilization Measurement and Management of Fleet Equipment

4 front loaders, graders, mechanical street sweeper trucks, air street sweeper trucks, sweepers/ scrubbers, riding mowers, truck tractors, snow removal attachments, rollers, drills, asphalt distributors, attachments, man lifts, and large trucks with a special body. Data Collection The research team conducted a national data collection effort and used data from seven state DOTs (California, Louisiana, Michigan, New Hampshire, Pennsylvania, Utah, and Washington) to develop utilization models and fleet management frameworks for 19 equipment types: dump trucks, pickup trucks, automobiles, vans, sport utility vehicles, trailers, front loader trucks, graders, mechanical street sweeper trucks, air street sweeper trucks, riding mowers, truck tractors, snow removal attachments, rollers, drills, asphalt distributors, attachments, man lifts, and large trucks with special body. Equipment Fleet Utilization Prediction The utilization measurement metrics used were “annual mileage,” “annual engine hours,” and “frequency of usage.” The review of the literature and the agency survey suggested that annual mileage was the most appropriate and recorded utilization metric for moving equipment types, and the research team collected a large amount of annual mileage data. The annual engine hours and frequency of usage were more appropriate for stationary equipment types and equipment types without an engine, respectively; the research team collected annual engine hours and frequency of usage data for some equipment types. The data were cleaned before modeling by removing outliers. The data were aggregated at the region level; the models estimated the annual mileage, annual engine hours, and frequency of usage for a specific equipment type in a region. Table 1 defines the parameters that were used in the prediction models. A comprehensive model fitting process was followed for each equipment type. The model considered various model structures including linear, quadratic, power, logarithmic, and nonlinear forms. Logarithmic linear regression, in which the independent variable is the logarithmic function of utilization metrics, yielded the best fit and the most intuitive model for predicting the annual mileage and engine hours for most equipment types. Table 2 lists all the equipment utilization estimation models. Equipment Fleet Utilization Management The research team developed an equipment utilization management framework for estimating the optimal utilization values and fleet composition in each region of a state. Variables Description : Annual Mileage Average annual mileage in a region : Annual Engine Hours Average annual engine hours in a region : Frequency of Usage Average frequency of usage in a region : Purchase Cost ($1,000) Average equipment purchase cost in a region : Annual Downtime Hours Average equipment annual downtime hours in a region : Annual Scheduled Maintenance Cost ($1,000) Average equipment annual scheduled maintenance cost in a region : Annual Unscheduled Repair Cost ($1,000) Average equipment annual unscheduled repair cost in a region : In-Service Age Average equipment in-service age in a region : Fleet Size Average fleet size in a region : Class 1: if the equipment is in class based on the NAFA class code Note: identifies equipment class as defined in the National Association of Fleet Administrators (NAFA) classification. Table 1. Utilization parameters used in utilization models.

5 Asset Type Utilization Estimation Model Dump Truck log( ) = 9.321 + 0.238 log( ) − 0.635 log( ) + 0.129 log( ⁄ )− 0.503 ( 5 + 6 + 7 ) Pickup Truck log( ) = 7.081 + 0.461 log( ) − 0.153 ( )0.5 + 0.275 log( ⁄ ) + 0.237 log( ) − 0.024 ( )0.5 − 0.049 2 − 0.144 3 − 0.464 4 Automobile log( ) = 9.119 + 0.256 + 0.0781 log( ) − 0.147 log( ) Van log( ) = 9.256 − 0.111 + 0.103 log( ⁄ ) + 0.188 log( )− 0.002 + 0.289 ( 5 + 6 ) Sport Utility Vehicle log( ) = 10.222 − 0.461 log( ) + 0.248 log( ) − 0.002 Trailer log( ) = 5.349 + 0.814 log( ) − 0.003 ( )2 + 0.663 log( ) + 0.481 log( ) − 0.009 = 135.65 + 1.365 + 18.073 log( ) − 125.681 12 + 61.966 14 + 46.051 17 Front Loader Truck log( ) = 5.590 − 0.004 ( )2 + 0.786 ( 3 + 4 ) Grader log( ) = 6.710 − 0.861 log( ) + 0.493 log( ) + 0.905 log( ) log( ) = 7.619 + 0.505 log( ) + 1.113 log( ) − 0.455 log( ) Mechanical / Air Street Sweeper Truck log( ) = 12.985 − 1.812 log( ) + 0.343 log( ⁄ ) − 0.470 log( ) − 0.096 log( ) + 0.973 4 log( ) = 6.291 − 0.220 + 0.541 log( ) + 0.416 log( ) Riding Mower log( ) = 4.445 + 0.222 log( ) + 0.440 log( )log( ) = −0.921 + 0.512 log( ) − 0.042 − 0.441 + 1.384 log( ) Truck Tractor log( ) = 10.014 − 0.004 ( × 10 −1)2 − 0.810 log( ) + 0.061 + 1.064 3 Snow Removal Attachment log( ) = −0.674 + 1.042 log( ) + 0.095 ( ⁄ ) − 0.015 Roller log( ) = 4.399 + 1.095 log( ) − 1.148 log( ) + 1.072 log( ) − 0.872 2 + 0.624 3 log( ) = 4.196 + 10.235 ( )−2 − 0.004 − 0.806 log( ) × log( ) + 2.084 2 log( ) = 129.893 + 0.316 ( ⁄ )2+107.621 3 Drill log( ) = 4.079 + 0.008 + 0.163 − 0.153 ( ) 0.5 = 26.481 + 16.841 ( ⁄ ) Asphalt Distributor log( ) = 22.406 − 2.417 log( ) − 0.476 log( ) − 0.005 ( ) 2 − 3.527 ( )2 Attachment log( ) = 5.056 + 0.354 log( ) − 0.127 ( )0.5−1.031 2 Man Lift log( ) = 4.101 + 0.938 log( ) + 0.317 − 0.167 ( ) 0.5 log( ) = 7.766 − 0.031 ( )2 + 1.653 × − 0.754 log( ) Large Truck with Special Body log( ) = 9.712 − 0.080 + 0.158 log( ) − 9.27 ( × 10−3)2 + 0.284 2 − 0.296 5 − 1.032 8 − 0.278 11 Table 2. Equipment utilization estimation models.

6 The framework incorporates a mathematical program that determines for each region in a specific year (a) fleet size, (b) number of equipment units to be purchased, (c) number of equipment units to be salvaged, (d) number of equipment units to be relocated to another region, and (e) average utilization levels required to meet the demand. The program determines the lowest total fleet management costs considering operating cost (i.e., the summation of annual fuel, unscheduled repair, and scheduled maintenance costs), equipment purchase cost, and equipment relocation cost. The program also updates the fleet size in each region in the following year considering the number of purchased equipment units, “brought-in” or “sent-out” equipment from/to other regions, and salvaged equipment. In addition, the program ensures that: 1. The number of equipment units leaving each region to other regions in the following year does not exceed the number currently available in that region; 2. The demand will be met by having sufficient fleet size in the following year (i.e., the product of fleet size and average utilization of each equipment type in each region in the following year shall be equal to or greater than the total demand); 3. No equipment will be utilized in excess of its maximum allowed utilization level; and 4. The number of purchased, salvaged, and relocated equipment units will not be less than zero for the following year. The mathematical model showed a reduction in total costs as a result of optimizing the fleet size; the number of equipment units to be purchased, salvaged, and relocated; and the region-level average utilization level for each equipment type. Guide for Equipment Fleet Utilization Measurement and Management A guide for utilization measurement and management of fleet equipment was developed as part of this research (Part II of this report). The guide incorporates a utilization manage- ment framework that integrates the (i) data input capabilities and (ii) utilization measure- ment models developed through rigorous statistical analyses of the collected data. These models use cost functions to determine the optimal utilization levels while considering operational constraints. Models were developed for different types of equipment assets and incorporated in a soft- ware application dubbed utilization prediction and management (UPM). The UPM soft- ware uses these models and the data available for the different equipment types to estimate the levels of utilization metrics and determine the optimal values for utilization measures (i.e., the values that minimize fleet management costs). Figure 1 shows snapshots of some steps in the UPM process. A user manual that details software installation, configuration, and capabilities is presented in Part III.

7 Figure 1. Snapshot of some steps in the UPM process.

8 Introduction 1.1 Background Fleet assets are critical for the delivery of state highway agency programs, projects, and services and consume a major portion of capital investments. These assets also require significant resources to ensure their operational perfor- mance, reliability, and service level. Agencies need to main- tain fleets of adequate sizes to meet the demand and avoid over- or underutilized equipment. Hence, there is a trade-off between minimizing resource investments and maintaining sufficiently-sized equipment fleets. Effective approaches to management of equipment fleet can help highway agencies meet the demand and extend equipment life expectancy by avoiding over- or underutilization, thus accruing significant cost reductions and environmental benefits. A major portion of fleet management activity is devoted to measuring, moni- toring, and reporting equipment utilization levels while meeting the needs of the highway agencies. The literature review revealed different definitions for utilization management. For example, it is defined as (a) management of the size and composition of the fleet [1], (b) an effort to balance under- and overutilized equipment and vehicles for more efficient and cost-effective use of the fleet assets [2], or (c) a process to decrease the total number of state-maintained vehicles to reduce the overall cost [3]. Further, asset management aims to identify the right type of equipment for each use and how each equipment item will be used. Different procedures are employed to address the utiliza- tion measurement and management challenges for highway agencies [4, 5, 6]. However, there is no widely accepted process for determining utilization criteria, measurement, and management of fleet equipment. Hence, a guide for utilization measurement and management of fleet equip- ment is needed. Such a guide needs to incorporate rational processes and suitable computer-based tools that provide optimal solutions to help equipment managers and admin- istrators make decisions regarding the fleet size required to meet agency needs. It should be recognized that utiliza- tion can only be managed for equipment with tracked and reported usage, and although a typical highway agency fleet includes many equipment types, only some of this equipment is suitable for tracking utilization. 1.2 Project Objective The objective of this research was to develop a guide for utilization measurement and management of fleet equip- ment and associated processes and electronic-based tools for use by state highway agencies. 1.3 Research Approach The project included a literature review of fleet utilization measurement and management, a survey to identify state DOTs’ current practices, analysis of available data and model development, and the preparation of the guide and utiliza- tion measurement and management tool. 1.4 Organization of the Report This report has three parts. Part I describes the research background (this chapter), research approach (Chapter 2), research findings and applications (Chapter 3), and conclu- sions and suggested research (Chapter 4). Part II is the guide for utilization measurement and management of fleet equip- ment, and Part III is the user manual for the developed UPM software. The software is available from the TRB website at trb.org by searching for “NCHRP Research Report 957.” C H A P T E R 1

9 2.1 Introduction Highway agencies use a variety of processes for utilization measurement and management of equipment fleet. However, there are no widely accepted processes for utilization mea- surement and management of fleet equipment. This research project was initiated to develop a guide that incorporates rational processes and appropriate electronic-based tools to address utilization measurement and management of fleet equipment. The research team identified and reviewed literature rele- vant to fleet utilization measurement and management and conducted a survey of state DOTs to identify fleet utiliza- tion measurement and management practices, utilization metrics, common equipment types and classes, and criteria used for utilization management. Based on the findings of the literature review and agency survey, the research team collected and processed data from several state DOTs to develop models for estimating fleet equipment utilization. Various mathematical forms for models and variable selec- tion approaches were considered to determine significant contributing factors to utilization. Several statistical tests were used to validate the developed models and ensure the validity of their assumptions. Also, models to estimate the operating cost of fleet equipment were developed following the same procedure. The equipment utilization and operating cost models were then incorporated in an optimization program to manage equipment utilization by determining the lowest total management cost while meeting agency needs and avoiding under- and overutilization of equipment. The program esti- mates optimal fleet size, utilization level, and the number of equipment units to be purchased, salvaged, and relocated in each region of a state; and was validated through sensitivity analyses. All utilization prediction and management models were then incorporated in UPM software. 2.2 Literature Review The literature review covered domestic and foreign litera- ture, research findings, and other information relative to utilization measurement and management of fleet equip- ment. The review focused on the practices, processes, and methodologies for the management of fleet equipment to help identify the most important utilization measurement metrics, factors influencing fleet utilization measurement and management, the common equipment types that highway agencies maintain, and utilization thresholds. 2.3 Agency Survey The survey of state DOTs identified fleet utilization measurement and management practices used by highway agencies, utilization measurement metrics, factors influenc- ing fleet utilization measurement and management, the most common equipment types that state DOTs maintain, and the types and amount of fleet utilization measurement and management data collected by highway agencies. 2.4 Data Collection The survey of state DOTs identified the type and amount of data recorded by state DOTs and their availability for use in this research. The survey indicated that the data collected included equipment identification number, NAFA class code, model year, manufacturer, report year, in-service age (year), fleet size, equipment capacity, fuel type, region, in-service date (year/month), ownership (own or rent), purchase cost ($), annual rent hours and cost, mounted equipment, annual mileage, annual engine hours, annual fuel consumption, frequency of usage per year (days scheduled), annual down- time hours, maintenance provider, annual (scheduled) C H A P T E R 2 Research Approach

10 maintenance hours and cost, (scheduled) maintenance interval mileage and days, seasonality, and predicted annual demand (number of miles, hours, days, etc.). The research team sought other data, such as annual cost data on (unscheduled) repair, insurance, interest charge, depreciation charge, taxes, licensing, storage, and overhead. 2.5 Data Processing To prepare the collected data for fitting the regression models, the influential outliers in a set of explanatory vari- ables were identified using Cook’s distance measure [7, 8]. This approach identifies the observations that negatively affect the regression model and removes them from the dataset. The approach considers a combination of each observation’s leverage and the associated residual values. Larger leverage and residual values yield a higher Cook’s distance and thus a higher chance that the observation is an outlier; observations with a Cook’s distance of more than three times the mean value were considered outliers and were removed from the analysis. In addition, only those variables having no more than 40% of their data missing were considered to ensure that sufficient data were avail- able to develop valid utilization estimation models. The collected data were aggregated both over state-defined regions and annually. 2.6 Fitting the Models To determine the most appropriate structure for develop- ing the regression models, linear, quadratic, power, logarith- mic, and nonlinear forms were considered; also, pairwise plots of the relationships between the equipment utilization metrics and the explanatory variable were constructed and evaluated for fitness. The most relevant explanatory variables were identified after the model structure was selected. Forward selection [9] and backward elimination [10] approaches were used to identify the best models; these were further improved by improving their fit using the R package [11]. The Pearson correlation and the variance inflation factor tests before and after fitting the model, respectively, were used to ensure that there was no multi- collinearity between the independent variables. 2.7 Validating the Models The validation process started with plotting the residual distribution. Residuals were defined as the difference between the predicted (by the model) and observed (from data) values of the dependent variable (e.g., annual mileage, annual engine hours, and frequency of usage). Thus, a value of zero means that the model exactly predicts the observation. Ideally, the residuals should follow a normal distribution with an average of zero (see Figure 2). After visual analysis of residual distributions, the Shapiro–Wilk test [12] was applied to each model to determine if the residuals of the regression model followed a normal distribution. In the next step, plots of predicted versus observed values for each model were prepared and a regression line was fitted. Ideally, the regression line should go through the origin and have a slope of one, with a high R-squared value. Figure 3 shows an example of a plot of predicted versus observed data, with a regression line having a slope of 0.999 and an R-squared value of 0.9. The Durbin–Watson statistic [13] was also used to inves- tigate autocorrelation in residuals (i.e., how residuals of subsequent observations in the dataset are correlated to each other, although the linear regression assumes that they are independent of each other). Another important assumption of the linear regression model is that there should be no heteroscedasticity of residuals (i.e., the variance of residuals should be constant and not increase as the fitted values of response variable increase); the Breusch–Pagan test [14] was used to evaluate this assumption. Figure 2. Example of residual histogram. Figure 3. Example plot of predicted versus observed data.

11 2.8 Developing the Equipment Fleet Utilization Management Program Determining the optimal fleet size and utilization level in a region requires consideration of budget, maximum utili- zation thresholds, and demand constraints while minimiz- ing the total fleet management costs. Determining how to achieve the optimal fleet size and utilization level in each region (by purchasing, salvaging, and relocating equipment) while minimizing total fleet management costs is a complex process that requires mathematical programming to account for all these constraints. The research team developed a math- ematical program for optimizing utilization level values and fleet composition in each region in each state. The program determines the following items for each year: 1. Fleet size in each region, 2. Number of equipment units to be purchased in each region, 3. Number of equipment units to be salvaged in each region, 4. Number of equipment units to be relocated from one region to another, and 5. Average utilization levels required to meet the demand in each region. The program finds the lowest total fleet management costs, including the operating cost (i.e., the summation of annual fuel, unscheduled repair, and scheduled maintenance costs), equipment purchase cost, and equipment relocation cost. The program updates the fleet size in each region in the following year based on the decisions regarding (a) purchased equip- ment, (b) “brought-in” or “sent-out” equipment from/to other regions, and (c) salvaged equipment and ensures that (see Figure 4): 1. The number of equipment units leaving each region to other regions in the following year does not exceed the number currently available in that region; 2. The demand will be met by having sufficient fleet size in the following year (i.e., the product of fleet size and average utilization of each equipment type in each region in the following year shall be equal to or greater than the total demand); 3. No equipment will be utilized in excess of its maximum allowed utilization level; and 4. The number of purchased, salvaged, and relocated equip- ment units will not be less than zero for the following year. The UPM software produces state-, region-, and unit-level results. 2.9 Validating the Equipment Fleet Utilization Management Program The utilization management program was validated through a rigorous sensitivity analysis of the results obtained from numerous input scenarios (e.g., combinations of various demand levels, relocation costs, and maximum utilization levels). Changes were made as necessary to ensure that the findings yield the expected results and follow the expected trends (e.g., increasing the demand level leads to an increase in the fleet size). 2.10 Developing the UPM Software The data input capabilities, prediction models, fleet equipment management program, and reporting capabilities were packaged in an electronic tool coded in Java that allows ready application by the end-user. 2.11 Validating the UPM Software Numerous input scenarios were developed for each feature of the software and used to validate the software and ensure all features work properly. * Purchase * Minimize costs - Fleet size - Utilization threshold * Relocate resources * Salvage Figure 4. Schematic overview of the utilization management program.

12 3.1 Introduction This chapter presents the findings of the literature review and agency survey; discusses the data collection, cleaning, and processing; and describes the development and valida- tion of the utilization estimation models, the fleet utilization management program, and the UPM software. 3.2 Literature Review The literature review focused on practices, processes, and methodologies as they relate to the utilization management of fleet equipment. The reviewed literature was categorized into five topics: (1) benefits, (2) fleet assignment, (3) environ- ment, (4) fleet maintenance, and (5) data collection. 3.2.1 Utilization Management Benefits The research team reviewed utilization management strategies used by transportation agencies in Ohio, the City of San Bernardino, Minnesota, Utah, California, North Carolina, Texas, and Pennsylvania. The utilization manage- ment approaches used by these agencies focus on (a) cost minimization (North Carolina, Pennsylvania, and Texas), (b) standardization of equipment categories (North Carolina, Ohio, and the City of San Bernardino), and (c) performance level determined by policy-based equipment monitoring (California, Minnesota, North Carolina, Ohio, Utah, and the City of San Bernardino). The Ohio Department of Transportation (ODOT) [15] adopted strategies aimed at improving fleet management program performance and operations, reducing associated costs, facilitating decision-making, and contributing to public accountability. The decision-making criteria for utili- zation management were based on cost minimization and benefit-cost analysis. ODOT categorized equipment based on their general value, widespread use across the state, and the number of pieces of equipment [15]. The main category included two separate groups of heavy equipment that is primarily used for construction and maintenance activities. One group included backhoes, excavators, graders, rollers, and crawler tractors; the other included construction brooms, bucket trucks, distributors, crawler and force-feed loaders, milling machines, asphalt reclaimers, seeders, and road wideners. Tractors, vocational vehicles, passenger cars, light dump trucks, brush chippers, skid steer loaders, front-end loaders, cleaner vacuums, and truck tractors were considered as separate categories. ODOT uses its own guidelines for minimum annual engine hours and mileage for different categories of equip- ment under seasonality constraints. These guidelines define the minimum utilization requirement for different types of equipment based on their usage in different seasons to help reduce the cost of owning/operating underutilized assets. Based on these guidelines, ODOT estimated that nearly 42% of their heavy equipment, 37% of tractors, 6% of voca- tional vehicles, 19% of passenger cars, 10% of light dump trucks, 73% of brush chippers, 41% of front-end loaders, 35% of cleaner vacuums, and 7% of truck tractors were used less than their expected operational time, and that having the right fleet size would save about $19,504,000 in a ten- year period [15]. To evaluate the condition of fleet equipment and each department’s permanent/temporary equipment needs, the City of San Bernardino adopted some guidelines for catego- rizing equipment and measuring utilization of the functional classes. The fleet comprised 728 units, including sedans, trucks, police vehicles, construction equipment, refuse trucks, and miscellaneous equipment. The vehicle and equipment fleets were categorized into seven functional categories under two broad classifications: “general use” and “special use” [16]. The utilization measures included monthly average mile/ hours to help identify overutilized/underutilized units for potential reassigning or disposal based on minimum C H A P T E R 3 Research Findings

13 utilization criteria. Fleet size reduction, makeup adjustment, and the number of pool units were considered based on an estimated salvage value, annual maintenance and repair cost savings, and replacement cost savings. Utilization thresholds were developed for high, medium, and low use (i.e., usage exceeding 80%, 50%–80%, and less than 50% of the annual average mileage/average monthly hours, respectively). The utilization data included units, city identification number, make, model, year, classification, department/ division assignment, domiciled location, current odometer and hour meter reading, total months in service, and usage over the last 12 months. Additionally, each unit’s intended use, equipment and loads transported, destinations, and frequency of use were determined through questionnaire responses. The analysis of the data resulted in the recom- mendation of a 6% reduction in the total number of examined city fleet units. The study also estimated that the fleet size adjustment would result in savings of about $3,516,520 over a ten-year period [16]. The Minnesota Department of Transportation (MnDOT) initiated use of fleet management performance measures in 2002 [17]. These measures target equipment utilization, units out-of-life cycle, fleet size, and scheduled versus unscheduled maintenance. MnDOT established a set of “equipment utilization rate goals” for mobile equipment including light, medium, and heavy-duty vehicles, snow plow trucks, loaders, mowers, tractors, and motor graders, among others. The minimum utilization requirements for different types of equipment were identified based on their usage in different seasons. The key factors that influenced equipment utili- zation and management strategies were the annual mileage, annual engine hours, and seasonality. The target minimum annual utilization was set at 8,000 miles per year (3,500 miles for seasonal snowplow trucks) and 500 hours per year (125 hours for seasonal snow and ice support equipment). These activities increased statewide utilization from 57% in 2002 to 73% in 2005. In 2006, MnDOT re-evaluated the utilization standard for different classes of equipment and introduced the following criteria: • 12,000 miles per year for most light-duty automotive classes; • 8,000 miles per year for most medium-/heavy-duty trucks; • 8,000 miles per year for tandem-axle snowplow trucks; • 6,000 miles per year for single-axle snowplow trucks; and • 500, 350, 250, or 125 hours per year for off-road heavy equipment and tractors. These utilization thresholds decreased the statewide utili- zation in 2010 to 61% [17]. Utah Department of Transportation (UDOT)’s Division of Fleet Operations maintains a vehicle information system that houses data on state-owned/operated motor vehicles. Based on minimum utilization standards, the state recom- mended strategies for underutilized equipment through a rotation program (e.g., selling equipment at optimal miles and years) and accrued savings in annual fleet turnover and maintenance costs [3]. Key factors in equipment utilization management included the mileage per month, frequency of use, and purpose/need for vehicles that did not meet the mileage and frequency-of-use criteria. The study of UDOT’s fleet suggested that, when feasible, renting seasonal and low utilization equipment would be cost-effective. The study also recommended central utilization management over the entire state and other activities to reduce the number of underutilized fleets, such as providing incentives to discourage agencies from retaining vehicles that are not used [3]. Caltrans used utilization, preventive maintenance, reten- tion, and availability/downtime for fleet management as fleet performance measures [18]. Caltrans’ division of research and innovation gathered information on budgeting methods and approaches for using utilization measures to reduce costs of fleet equipment management procedures from six state DOTs (Illinois, New York, North Carolina, Pennsylvania, Texas, and Virginia); four of which (Illinois, New York, North Carolina, and Pennsylvania) use computer-based systems to collect the data. The most influential factors in allocat- ing budget for equipment replacement are equipment hours used, mileage, repair cost, and downtime. These six DOTs usually spend $20 to $60 million annually to replace fleet equipment; all except North Carolina noted serious equip- ment backlog problems and often reduced fleet equipment management costs by adopting fuel efficiency policies, fleet size reduction and reliance on equipment leasing, and reduc- tion in services. Kauffmann et al. introduced a method for fleet manage- ment performance monitoring to enhance and expand the analytical models developed by the North Carolina Depart- ment of Transportation for equipment usage, cost, and optimal life cycle [19]. Another study developed economic analysis models that reflected utilization for different classes of equipment and calculated ownership, operations, and equipment life cycle costs [20]. This study suggested that utilization by itself may not truly reflect the usability and performance of equipment and recommended average operating rate (cost/mile or cost/hour) as the utilization metric; it also analyzed equipment based on annual costs when data on equipment usage were unavailable. Exponen- tial regression models were used to measure the decline of an equipment item’s market value. To estimate the operating cost, an equation of the relationship between the average annual operating rate and equipment age was fitted for each class using the least square approach; it showed that the average annual operating rate increased when equipment

14 age increased. Also, a linear regression model for annual equipment use versus age showed the decline in usage over time. Two cost models that estimate the economic life for each class of equipment (i.e., the period at which the average total cost rate reaches a minimum value) were considered; these are the equivalent uniform annual cost and the eco- nomic life model based on the present value of the average life-to-date rate (cost/mile or cost/hour). The study, which included more than 20 equipment types, showed that hours or miles per year and cost/mile or cost/hour were the most important factors to consider in equipment utilization and management strategies. It also provided a consistent means for managing fleet equipment, especially for equipment with less variability in annual usage. The Texas Department of Transportation (TxDOT) con- ducted a performance measure preventive maintenance study for a single category of equipment [21]. The TxDOT used a program (named FleetTrackS) to track data collected over time by an onboard diagnostic system for vehicle miles or operational hours. A stepwise regression method was used to relate oil degradation levels to operational engine data and identify the appropriate time for specific mainte- nance needs (e.g., oil change interval). The study estimated that using the optimal oil change interval would save about $16,000 of maintenance cost annually for a single equip- ment type. The Pennsylvania Department of Transportation (PennDOT) used a Macro software tool based on statistical models designed to predict the fleet’s life cycle [22]. The goal was to predict the probability of repairs/replacements of fleet components and the cost of maintenance and repairs for each equipment type. The software prioritizes equip- ment replacement based on their cost and age to minimize total fleet management costs and maximize the readiness of equipment fleet. PennDOT gathered data from the SAP plant maintenance program over a five-year period (2007 to 2012) for different types of equipment (dump trucks, excavators, front-end-loaders, backhoes, and crew-cabs). The data on (a) number of pieces of each equipment type, (b) total number of maintenance/repair records for each type, (c) average cost per record, (d) minimum and maxi- mum costs observed in these records, and (e) total costs of maintenance and repairs summed across all records were used for the analysis. In this study, three utilization factors (fuel usage, personnel hours, and monthly equipment age) were used to predict the monthly cost for each type of equipment. The plot of cost ratio (cumulative cost/fuel usage or cumulative cost/personnel hours) versus equip- ment age (in years) helped determine the life cycle: the age where a substantial change in the slope of the graph is observed. 3.2.2 Utilization Management and Fleet Assignment Resource allocation can affect fleet utilization; costs can be reduced by properly allocating equipment to different regions or time of usage. To improve utilization and reduce the fleet size, MnDOT suggested using assignment measures to allocate equipment to different regions based on statistical data [23]. A graphical information system was also recom- mended to monitor fleet performance metrics and improve sharing information between districts to help assign the right fleet size to each district. MnDOT also suggested a central statewide utilization management strategy to achieve the highest performance. Lee et al. used models for the optimal assignment of truckloads to the delivery time to improve fleet utilization (e.g., by shifting load start-times) [24]. These models helped determine the minimum required fleet size under tardiness/ earliness costs (i.e., time window constraints). The models also showed that improvement in fleet utilization would significantly reduce the number of equipment units in use. Regan et al. developed a simulation framework using standard Monte Carlo techniques for dynamic fleet manage- ment to evaluate the impact of some changes (e.g., stochastic demand, driver and fleet availability, and traffic condition) on real-time fleet operation while maintaining the level of service and profitability [25]. The results indicated that lower fleet utilization was associated with higher operational costs and lower profits. For example, the operational costs per mile for a fleet that worked at 50% utilization were about 19% higher than for a fleet that worked at 100% utilization. 3.2.3 Utilization Management and Environment Figliozzi et al. analyzed the impact of utilization (mileage per year per vehicle) and gasoline prices on vehicle-purchasing decisions and developed a replacement model based on fleet costs and utilization [26]. The model also considered green- house gas (GHG) costs (e.g., GHG equivalent life cycle costs) for multiple vehicle types, thus integrating the traditional fleet management costs with environmental effects. The study indicated that fuel-efficient vehicles such as hybrids and electric vehicles should be purchased only when the gasoline prices or fleet utilization were high. Jin and Kite- Powell [27], Chen and Lin [28], and Lin et al. [29] used statistical analyses of fleet data (e.g., age, annual mileage, utilization level, mile-per-gallon rate, purchase cost, opera- tions and maintenance cost, salvage value, and emission cost) to examine the impact of utilization, policy, market, and environment and technological factors on fleet management. These studies revealed the sensitivity of fleet utilization and

15 replacement strategies to fuel cost (i.e., future change of the fuel cost would influence long-term planning), and suggested multi-stage planning for fleet management strategies. 3.2.4 Utilization Management and Fleet Maintenance Fleet managers encounter operational constraints and sources of uncertainty (e.g., fleet breakdown) in large fleet maintenance scenarios. Decision-making under these com- plex conditions can be facilitated using optimization methods. For example, a controlled maintenance management system can help identify the most efficient utilization of labor and equipment [30]. Lee et al. indicated that using predictive tools to monitor degradation would improve utilization and increase the cost-benefits by up to 69% compared to repair- ing after failure [31]. Using survey results and field data, Vujanović et al. [32] evaluated management indicators used for the maintenance process, the transport process, and the environment. The important indicators in efficient fleet maintenance management were maintenance plan realization (most important), operational plan realization percentage (second rank), and vehicle payload utilization (third rank). 3.2.5 Utilization Data Collection Tracking equipment utilization helps manage equip- ment usage levels and fleet efficiently. Several approaches are available for collecting the data required for measuring equipment utilization. Said et al. proposed an algorithm to extract fleet usage data based on GPS information for equip- ment location and time of day [33]. The New York State DOT used an onboard data logging system to capture the equip- ment’s daily operational data including engine parameters and global positioning information to prevent excessive idling of their equipment [34]. The data showed that idling occurs often between 9 AM and 7 PM for ¾ ton pickup trucks, between 11 AM and 7 PM for passenger pickups, and at midday for stake rack trucks. Although gathering data from onboard units helps understand if fleets are used efficiently, predicting utilization is needed for future planning; statistical or machine learning analysis could be used. For example, Kargul et al. used the support vector machine algorithm to predict the utilization of heavy equipment [35]—an approach that showed a good approximation of the real utilization rate. 3.2.6 Summary of the Literature Review Findings The literature identified the parameters used to measure equipment utilization and the factors influencing utilization; these are listed below. 1. Parameters used for measuring equipment utilization: – Annual mileage, – Annual engine hours, – Usage over the last 12 months, – Seasonality, and – Frequency of use. 2. Factors that influence equipment utilization (directly or indirectly): – Equipment in-service age, – Cumulative utilization level, – Breakdown rate, – Downtime hours, – Fleet size, – Number of equipment under repair, – Initial workload of equipment, – Agency management policies and practices, – Normal equipment life, – Expected equipment workload, – Product types, – Environmental impacts, – Site condition uncertainty, – Purchase cost, – Operating cost, – Maintenance cost, – Repair cost, and – Salvage value. 3.3 Agency Survey A survey of state DOTs (and that of the District of Columbia) provided information on the state of practice in utilization measurement and management of fleet equipment. The survey aimed at identifying (a) highway fleet equipment types used (own/rent/lease), (b) factors considered in utili- zation measurement practices, and (c) methodologies used for utilization measurement practices. 3.3.1 Fleet Equipment Type and Classification The survey listed 28 equipment types based on the NAFA classification system and requested additions if necessary. Based on the responses received, the most commonly used 20 equipment types were identified and listed in Table 3. 3.3.2 Utilization Measurement Metrics The utilization metrics most commonly used by state DOTs were annual mileage, annual engine hours, usage over the last 12 months, and frequency of use. The annual mileage can only be used for moving equipment, and annual

16 # Equipment Class NAFA Class Code 1 Dump Truck 8,501-10,000 GVW 2712 10,001-14,000 GVW 3712 14,001-16,000 GVW 4712 16,001-19,500 GVW 5712 19,501-26,000 GVW 6712 26,001-33,000 GVW 7712 > 33,000 GVW 8712 2 Pickup Truck < 8,500 GVW 1510, 1511, 1512, 1513, 1514, 1520, 1521, 1522, 1523, 1524, 1530, 1531, 1532, 1533, 1534, 1540, 1547, 1548 10,001-14,000 GVW 3510, 3511, 3512, 3513, 3514, 3540, 3547, 3548 14,001-16,000 GVW 4510, 4511, 4512, 4513, 4514, 4520, 4527, 4528 8,501-10,000 GVW 2510, 2511, 2512, 2513, 2514, 2540, 2547, 2548 3 Automobile < 8,500 GVW 1310, 1311, 1312, 1313, 1320, 1321, 1322, 1323, 1324, 1330, 1331, 1332, 1333, 1338, 1340, 1341, 1342, 1343, 1348 4 Van < 8,500 GVW 1410, 1411, 1412, 1413, 1414, 1418, 1420, 1421, 1422, 1423, 1424, 1425, 1426, 1427, 1428 10,001-14,000 GVW 3410, 3411, 3412, 3413, 3414, 3420, 3421, 3422, 3423, 3424, 3425, 3426, 3427, 3429 14,001-16,000 GVW 4410, 4411, 4412, 4413, 4414, 4420, 4421, 4422, 4423, 4424, 4425, 4426, 4427, 4429 16,001-19,500 GVW 5420, 5421, 5422, 5423, 5424, 5425, 5426, 5427, 5428, 5429 19,501-26,000 GVW 6420, 6421, 6422, 6423, 6424, 6425, 6426, 6427, 6428, 6429 8,501-10,000 GVW 2410, 2411, 2412, 2413, 2414, 2420, 2421, 2422, 2423, 2424, 2425, 2426, 2427, 2429 5 Sport Utility < 8,500 GVW 1610, 1611, 1612, 1620, 1621, 1622, 1623, 1630, 1631, 1632, 1633, 1634, 1640, 1647, 1648 10,001-14,000 GVW 3610, 3611, 3612, 3613, 3614, 3640, 3647, 3648 8,501-10,000 GVW 2610, 2611, 2612, 2613, 2614, 2640, 2647, 2648 6 Trailer Air Compressor 320, 321, 322, 323 Auxiliary Power 390, 397, 398, 399 Boat 780, 781, 782, 783 Concrete Mixer 380, 381, 382, 383 Construction 0360 Dump Body 750, 751, 752, 753 Flat Bed 710, 711, 712, 713 Generator 310, 311, 312, 313, 314, 315, 316, 317, 318 Office 790, 791, 792, 793 Pressure Washer 340 Public Utility 350, 351 Pump 330, 331, 332, 333, 334 Refrigerator 760, 761, 762, 763 Sanitation 770, 771, 772, 773 Sewer Equipment 370, 371, 372, 373, 374 Tank Body 730, 731, 732, 733 Tilt Bed 720, 721, 722, 723 Van Body 740, 741, 742, 743 7 Front Loader 16,001-19,500 GVW 5763 19,501-26,000 GVW 6763 26,001-33,000 GVW 7763 > 33,000 GVW 8763 8 Grader Off-Road and Construction 9160 9 Mechanical Street Sweeper Truck 16,001-19,500 GVW 5771 19,501-26,000 GVW 6771 26,001-33,000 GVW 7771 > 33,000 GVW 8771 10 Air Street Sweeper Truck 16,001-19,500 GVW 5772 19,501-26,000 GVW 6772 26,001-33,000 GVW 7772 > 33,000 GVW 8772 11 Sweeper/ Scrubber Walk-Behind 9411 Riding 9412 Three-Wheeled 9413 12 Riding Mower Off-Road and Construction 9610, 9611, 9612 Table 3. Equipment commonly used by state DOTs.

17 engine hours can be used for equipment that has an engine, while other measurement metrics can be used for all types of equipment. Survey responses showed that, of 32 respond- ing state DOTs, 23 use “driven miles” as the main utilization metric. The respondents also noted that mileage metering is accurate, is easy to report, indicates actual usage, makes replacement predictable, and is consistent across various classes of an equipment type. But, it does not report equip- ment conditions and is hard to collect in equipment without an odometer. The survey also indicated that 20 state DOTs use “equip- ment age” and “engine hours.” The respondents also noted that the engine hours can provide the actual utilization of equipment and is easy to use, but it does not report equip- ment conditions and is confusing to collect for equipment having two engines. Frequency of use was noted for station- ary equipment without an engine. 3.3.3 Factors Contributing to Equipment Utilization The factors influencing equipment utilization identified in the survey responses are listed in Table 4; these factors are further discussed in Section 3.5. 3.4 Data Collection 3.4.1 Data Availability The survey of state DOTs identified the type and amount of data recorded by state DOTs and their availability for use in this research. Responses were received from 19 DOTs (Indiana, South Carolina, Oregon, Utah, Maryland, Ohio, Wyoming, California, Washington, Minnesota, Massachu- setts, South Dakota, Florida, Montana, Idaho, Colorado, Georgia, Alaska, and Oklahoma); all respondents indicated # Equipment Class NAFA Class Code 14 Snow Removal Attachment Nose Plows 111 V Plows 112 Wing Plows 113 Underbody Plows 114 De-icer Equipment 115 Snow Blowers 116 Material Spreaders 117 15 Roller Static 9441 Vibratory 9442 Compactor 9443 16 Drill Off-Road and Construction 9530 17 Asphalt Distributor 16,001-19,500 GVW 5778 19,501-26,000 GVW 6778 26,001-33,000 GVW 7778 > 33,000 GVW 8778 18 Attachment Spreader 0121 Aerator 0122 Soil Preparation 0123 Planter 124 Shredder/Mulcher 125 Mower 126 Bucket 131 Backhoe 132 Breaker 133 Tamper 134 19 Man Lift Off-Road and Construction 9330 20 Large Truck with Special Body Note: GVW = gross vehicle weight. Beverage Body 2716, 3715, 4715, 5715, 6715, 7715, 8715 Crew Cab 4513 Extended Cab 4512 Fifth Wheel 4718, 5718, 6718, 7718, 8718 Flat Bed 2711, 3711, 4711, 5711, 6711, 7711, 8711 Refrigerator Body 2717, 3716, 4716, 5716, 6716, 7716, 8716 Regular Cab 4511 Tanker 2718, 3717, 4717, 5717, 6717, 7717, 8717 Tilt Bed 2714 Utility Bed 2713, 3713, 4514, 4713, 5713, 6713, 7713, 8713 Van Body 2715, 3714, 4714, 5714, 6714, 7714, 8714 13 Truck Tractor 19,501-26,000 GVW 6810, 6820, 6830, 6840, 6890 26,001-33,000 GVW 7810, 7820, 7830, 7840, 7890 > 33,000 GVW 8810, 8820, 8830, 8840, 8890 Table 3. (Continued).

18 availability of the data for this research. Seventeen of these state DOTs indicated that the 20 equipment types listed in Table 3 represent the equipment types used by the DOT. Reported data availability is listed in Table 5. Each respond- ing state DOT recorded at least two years of data, and all respondents noted the availability of the data for use in this research. The survey indicated that 10 of 19 responding state DOTs did not have data on rent reason, annual rent hours, rent costs, annual idle hours, annual interest charge, annual taxes, annual licensing fee, annual storage cost, and estimated annual demand (e.g., number of miles, hours, or days). Also, six of the responding state DOTs had data on equipment capacity, annual downtime hours, seasonality, frequency of usage per year (days scheduled), and “other” annual opera- tion cost. 3.4.2 Data Requested and Used A request for data was distributed to the 19 state DOTs that indicated availability of data for use in this research (and subsequently to the remaining state DOTs). The request included definitions of the requested data elements and a template of the desired format for recording the data but noted that data may be provided in other formats. Data received from seven state DOTs [California (Caltrans), Louisiana (LaDOT), Michigan (MDOT), New Hampshire (NHDOT), Pennsylvania (PennDOT), Utah (UDOT), and Washington (WSDOT)] were used in this research. 3.4.3 Data Processing Data received were processed and analyzed using SAS 9.3 to determine missing and non-missing frequency and respec- tive percentages for each variable included in the dataset. The Excel files provided by state DOTs were read into SAS 9.3 data tables and presented as categorical values (e.g., equip- ment number, NAFA class code, and equipment descrip- tion) or numeric values (e.g., purchase cost, annual mileage, and annual downtime hours); missing data were noted. Reported zero values of some fleet utilization variables (e.g., annual mileage, annual fuel cost, and annual downtime hours) could be the result of not using the equipment in the reported year or not properly reporting the usage. The dataset was reviewed and obvious errors (e.g., entering a negative number for positive variables or a character entry for a numerical variable) were identified. The data were found to be adequate to build utilization prediction models for the following 19 equipment types: 1. Dump trucks, 2. Pickup trucks, 3. Automobiles, 4. Vans, 5. Sport utility vehicles, 6. Trailers, 7. Front loader trucks, 8. Graders, 9. Mechanical street sweeper trucks, Factor Description In-Service Age Years or month in service Fleet Size Number of equipment units in the agency with the same class code in the report year Equipment Capacity Measurement type used to define the equipment's capacity: pounds, gallons, tons County County where the asset is used the most in the report year In-Service Year/Month Year/month when the agency started using the asset (sometimes entered as the purchase delivery date) Ownership/Rental Own or rent Purchase Cost Purchase cost of the asset Annual Rent Hours Total hours the asset was rented in the report year Rental Cost Annual rent cost in the report year Rent Reason Seasonal use, unplanned need, extraordinary weather event, etc. Mounted Equipment E.g., dump body, aerial bucket, snowplow Total Mounted Equipment Cost Purchase cost of all equipment mounted on this asset (entered as "0" if there is no mounted equipment) Annual Downtime Hours Total hours per report year (out of scheduled hours) that equipment is unavailable Annual Maintenance Hours Total hours spent on maintaining the asset in the report year Annual Maintenance Cost Annual maintenance cost in the report year Maintenance Interval Mileage Meter reading between scheduled maintenance services Maintenance Interval Date Date of last scheduled maintenance services Annual Insurance Cost Annual insurance cost in the report year Annual Operation Cost Annual operating cost in the report year Annual Unscheduled Repair Costs Annual unscheduled repair cost in the report year Seasonality Seasons when the asset is used in the report year Annual Demand Predicted annual demand for the asset Table 4. Factors expected to influence equipment utilization.

19 10. Air street sweeper trucks, 11. Riding mowers, 12. Truck tractors, 13. Snow removal attachments, 14. Rollers, 15. Drills, 16. Asphalt distributors, 17. Attachments, 18. Man lifts, and 19. Large trucks with a special body. 3.5 Utilization Estimation Models The utilization models estimate the average annual mileage, annual engine hours, or frequency of usage for different equip- ment types and classes at the region level based on the data provided by state DOTs. Based on the literature review and agency survey, seven factors were found to affect utilization metrics statistically; these factors together with the utilization measurement metrics are listed and defined in Table 6. Table 7 further defines class variables for the different equipment types. Data Elements Data Available No Data Available Some Data Available Total Count Count Count Count Model year 19 0 0 19 Manufacturer 19 0 0 19 Report year 14 1 3 18 In-service age (year) 18 1 0 19 Fleet size (number of equipment units) 19 0 0 19 Equipment capacity 4 4 8 16 Fuel type 16 1 2 19 County 10 4 5 19 In-service date (year/month) 18 0 1 19 Ownership (own or rent) 18 1 0 19 Purchase cost ($) 17 0 2 19 Annual rent hours 2 10 5 17 Rental cost ($) 3 8 6 17 Rent reason 0 13 4 17 Mounted equipment (any "major" mounted equipment) 12 2 4 18 Total mounted equipment cost ($) 10 2 5 17 Annual mileage 16 1 2 19 Annual engine hours 12 4 3 19 Annual fuel consumption (gallons, GGE, kW-h, etc.) 12 3 3 18 Annual fuel cost ($) 12 3 3 18 Frequency of usage per year (days scheduled) 4 5 9 18 Annual downtime hours 5 8 5 18 Annual idle hours 1 13 4 18 Maintenance provider (in-house, outsourced, or both) 14 2 2 18 Annual (scheduled) maintenance hours (not included in "downtime hours") 7 6 5 18 Annual (scheduled) maintenance cost ($) 14 1 4 19 (Scheduled) maintenance interval mileage 15 1 3 19 (Scheduled) maintenance interval (days) 12 4 3 19 Annual (unscheduled) repair cost ($) 12 1 5 18 Annual insurance cost ($) 6 5 7 18 Annual interest charge ($) 2 13 3 18 Annual depreciation charge ($) 11 5 3 19 Annual taxes 3 13 2 18 Annual licensing fee ($) 4 11 3 18 "Other" annual operation cost ($) (excluding maintenance, repair, fuel, insurance) 5 5 7 17 Annual storage cost ($) 1 16 1 18 Annual overhead (indirect) cost ($) 6 7 4 17 Seasonality 5 8 4 17 Estimated annual demand (number of miles, hours, days, etc.) 2 11 5 18 Note: GGE = gasoline gallon equivalent. Table 5. Reported data availability.

20 Factor Description : Annual Mileage Average annual mileage in a region : Annual Engine Hours Average annual engine hours in a region : Frequency of Usage Average frequency of usage in a region : Purchase Cost ($1,000) Average equipment purchase cost in a region : Annual Downtime Hours Average equipment annual downtime hours in a region : Annual Scheduled Maintenance Cost ($1,000) Average equipment annual scheduled maintenance cost in a region : Annual Unscheduled Repair Cost ($1,000) Average equipment annual unscheduled repair cost in a region : In-Service Age Average equipment in-service age in a region : Fleet Size Average fleet size in a region Class 1: if the equipment is in class based on the NAFA class code Note: indicates equipment class as defined in the NAFA classification. Table 6. Definition of parameters used in utilization models. Asset Type Variable Description Dump Truck 1 Equipment weight between 8,501 GVW and 10,000 GVW 2 Equipment weight between 10,001 GVW and 14,000 GVW 3 Equipment weight between 14,001 GVW and 16,000 GVW 4 Equipment weight between 16,001 GVW and 19,500 GVW 5 Equipment weight between 19,501 GVW and 26,000 GVW 6 Equipment weight between 26,001 GVW and 33,000 GVW 7 Equipment weight greater than 33,000 GVW Pickup Truck 1 Equipment weight less than 8,500 GVW 2 Equipment weight between 8,501 GVW and 10,000 GVW 3 Equipment weight between 10,001 GVW and 14,000 GVW 4 Equipment weight between 14,001 GVW and 16,000 GVW 5 Equipment weight greater than 16,000 GVW Automobile 1 Equipment weight less than 8,500 GVW Van 1 Equipment weight less than 8,500 GVW 2 Equipment weight between 8,501 GVW and 10,000 GVW 3 Equipment weight between 10,001 GVW and 14,000 GVW 4 Equipment weight between 14,001 GVW and 16,000 GVW 5 Equipment weight between 16,001 GVW and 19,500 GVW 6 Equipment weight between 19,501 GVW and 26,000 GVW Sport Utility Vehicle 1 Equipment weight less than 8,500 GVW 2 Equipment weight between 8,501 GVW and 10,000 GVW Equipment weight between 10,001 GVW and 14,000 GVW 3 Trailer 1 Air Compressor 2 Auxiliary Power 3 Boat 4 Concrete Mixer 5 Construction 6 Dump Body 7 Flat Bed 8 Generator 9 Office 10 Pressure Washer 11 Public Utility 12 Pump 13 Refrigerator 14 Sanitation 15 Sewer Equipment 16 Tank Body 17 Tilt Bed 18 Van Body Front Loader Truck 1 Equipment weight between 16,001 GVW and 19,500 GVW 2 Equipment weight between 19,501 GVW and 26,000 GVW 3 Equipment weight between 26,001 GVW and 33,000 GVW 4 Equipment weight greater than 33,000 GVW Grader 1 Off-Road and Construction Mechanical/Air Street Sweeper Truck 1 Equipment weight between 16,001 GVW and 19,500 GVW 2 Equipment weight between 19,501 GVW and 26,000 GVW 3 Equipment weight between 26,001 GVW and 33,000 GVW Equipment weight greater than 33,000 GVW 4 Riding Mower 1 Off-Road and Construction Table 7. Class variables for different equipment types.

21 Table 8 lists the average annual mileage, annual engine hours, or frequency of usage estimation models for the 19 equipment types considered in this project. The table also shows the adjusted R-squared values and sample sizes; a high value indicates a good fit. The sample size in all cases was large enough to aggregate the data into a regional level and estimate the utilization predictions for each region in a state. As the models show, none of the explanatory variables were linearly correlated. Note that the same utilization esti- mation model applies to mechanical street sweeper trucks and air street sweeper trucks. 3.6 Equipment Cost Estimation Models Models for estimating the total operating cost of 18 equip- ment types were developed for use in fleet equipment manage- ment. The annual operating cost is defined as the summation of the annual fuel cost, unscheduled repair cost, and sched- uled maintenance cost, and it is the dependent variable in the regression analysis. The independent variables are annual mileage, annual downtime hours, in-service age, fleet size in a region, and class of equipment. Table 9 lists the cost functions (the variables are explained in Table 6). 3.7 Validating the Models Examination of the data indicated that (a) residual (differ- ences between the estimated and observed annual mileage) histograms for the utilization estimation models followed a bell-shaped curve around the value of zero; (b) the resid- uals are almost normally distributed; and (c) the regression models accurately estimate the actual average annual mileage in regions. Asset Type Variable Description Asphalt Distributor 1 Equipment weight between 16,001 GVW and 19,500 GVW 2 Equipment weight between 19,501 GVW and 26,000 GVW 3 Equipment weight between 26,001 GVW and 33,000 GVW 4 Equipment weight greater than 33,000 GVW Attachment 1 Spreader 2 Aerator 3 Soil Preparation 4 Planter 5 Shredder/Mulcher 6 Mower 7 Bucket 8 Backhoe 9 Breaker 10 Tamper Man Lift 1 Off-Road and Construction Large Truck with Special Body 1 Beverage Body 2 Crew Cab 3 Extended Cab 4 Fifth Wheel 5 Flat Bed 6 Refrigerator Body 7 Regular Cab 8 Tanker 9 Tilt Bed 10 Utility Bed Van Body Truck Tractor 1 Equipment weight between 19,501 GVW and 26,000 GVW 2 Equipment weight between 26,001 GVW and 33,000 GVW 3 Equipment weight greater than 33,000 GVW Snow Removal Attachment 1 Nose Plows 2 V Plows 3 Wing Plows 4 Underbody Plows 5 De-icer Equipment 6 Snow Blowers 7 Material Spreaders Roller 1 Static 2 Vibratory 3 Compactor Drill Off-Road and Construction Table 7. (Continued).

22 Table 8. Equipment utilization estimation models. Man Lift Annual Mileage log( ) = 4.101 + 0.938 log( ) + 0.317 − 0.167 ( )0.5 0.77 18/164 Engine Hours log( ) = 7.766 − 0.031 ( )2 + 1.653 − 0.754 log( ) 0.59 6/85 Large Truck with Special Body Annual Mileage log( ) = 9.712 − 0.080 + 0.158 log( ) − 9.27 ( × 10−3)2 + 0.284 2 − 0.296 5 − 0.86 76/12,690 Asset Type Utilization Metric Utilization Estimation Model Adjusted R- Squared Sample Size After/Before Aggregation Dump Truck Annual Mileage log( ) = 9.321 + 0.238 log( ) − 0.635 log( ) + 0.129 log( ⁄ ) − 0.503 ( 5 + 6 + 7 ) 0.65 98/15,338 Pickup Truck Annual Mileage log( ) = 7.081 + 0.461 log( ) − 0.153 ( )0.5 + 0.275 log( ⁄ ) + 0.237 log( ) − 0.024 ( )0.5 − 0.049 2 − 0.144 3 − 0.464 4 0.89 78/15,334 Automobile Annual Mileage log( ) = 9.119 + 0.256 + 0.0781 log( ) − 0.147 log( ) 0.74 32/4,327 Van Annual Mileage log( ) = 9.256 − 0.111 + 0.103 log( ⁄ ) + 0.188 log( ) − 0.002 + 0.289 ( 5 + 6 ) 0.80 55/3,384 Sport Utility Vehicle Annual Mileage log( ) = 10.222 − 0.461 log( ) + 0.248 log( ) − 0.002 0.81 37/2,956 Trailer Annual Mileage log( ) = 5.349 + 0.814 log( ) − 0.003 ( )2 + 0.663 log( ) + 0.481 log( ) − 0.009 0.68 72/1,273 Frequency of Usage = 135.65 + 1.365 + 18.073 log( ) − 125.681 12 + 61.966 14 + 46.051 17 0.67 39/3,392 Front Loader Truck Annual Mileage log( ) = 5.590 − 0.004 ( ) 2 + 0.786 ( 3 + 4 ) 0.83 17/170 Grader Annual Mileage log( ) = 6.710 − 0.861 log( ) + 0.493 log( ) + 0.905 log( ) 0.86 24/751 Engine Hours log( ) = 7.619 + 0.505 log( ) + 1.113 log( ) − 0.455 log( ) 0.87 6/111 Mechanical Sweeper Truck Annual Mileage log( ) = 12.985 − 1.812 log( ) + 0.343 log( ⁄ ) − 0.470 log( ) − 0.096 log( ) + 0.973 4 0.92 29/592 Engine Hours log( ) = 6.291 − 0.220 + 0.541 log( ) + 0.416 log( ) 0.75 19/459 Air Street Sweeper Truck Annual Mileage log( ) = 12.985 − 1.812 log( ) + 0.343 log( ⁄ ) − 0.470 log( ) − 0.096 log( ) + 0.973 4 0.92 29/592 Engine Hours log( ) = 6.291 − 0.220 + 0.541 log( ) + 0.416 log( ) 0.75 19/459 Riding Mower Annual Mileage log( ) = 4.445 + 0.222 log( ) + 0.440 log( ) 0.60 15/226 Frequency of Usage log( ) = −0.921 + 0.512 log( ) − 0.042 − 0.441 + 1.384 log( ) 0.89 12/516 Truck Tractor Annual Mileage log( ) = 10.014 − 0.004 ( × 10−1)2 − 0.810 log( ) + 0.061 + 1.064 3 0.77 31/483 Snow Removal Attachment Annual Mileage log( ) = −0.674 + 1.042 log( ) + 0.095 ( ⁄ ) − 0.015 0.74 32/715 Roller Annual Mileage log( ) = 4.399 + 1.095 log( ) − 1.148 log( ) + 1.072 log( ) − 0.872 2 + 0.624 3 0.91 25/259 Engine Hours log( ) = 4.196 + 10.235 ( )−2 − 0.004 − 0.806 log( ) log( ) + 2.084 2 0.58 11/61 Frequency of Usage log( ) = 129.893 + 0.316 ( ⁄ )2+107.621 3 0.74 18/137 Drill Annual Mileage log( ) = 4.079 + 0.008 + 0.163 − 0.153 ( )0.5 0.99 6/54 Engine Hours = 26.481 + 16.841 ( ⁄ ) 0.75 5/38 Asphalt Distributor Annual Mileage log( ) = 22.406 − 2.417 log( ) − 0.476 log( ) − 0.005 ( )2 − 3.527 ( )2 0.75 13/342 Attachment Annual Mileage log( ) = 5.056 + 0.354 log( ) − 0.127 ( )0.5−1.031 2 0.75 18/282

23 Sport Utility Vehicles = 0.24 − 1115.04 log( ) + 2679.59 log( ) − 3.99 0.81 19/538 Trailers = 0.002 ( )2 + 1300.84 1 + 2264.87 3 + 1286.74 5 + 399.71 6 + 1420.31 7 + 718.24 8 + 7828.37 9 + 873.75 10 + 1383.05 11 + 1315.92 12 + 2005.43 14 + 1699.40 16 + 633.10 17 0.72 89/343 Front Loader Trucks = 0.048 ( )2 + 2818.14 ( 1 + 2 ) + 8296.82 ( 3 + 4 ) 0.77 16/152 Graders = 359.76 ( ⁄ ) − 21.43 0.63 9/113 Mechanical/Air Street Sweeper Trucks = 19.87 + 51.72 ( )2 0.58 18/441 Riding Mowers = 2.87 − 8274.57 ( ⁄ ) 0.72 15/130 Truck Tractors = 0.92 − 25.349 + 3093.94 2 + 6933.45 3 0.74 26/44 Snow Removal Attachments = 11.4 × 10−5 ( 2 )⁄ − 1505.08 ( 2 )⁄ + 8266.77 1 + 3007.77 6 0.58 14/54 Rollers = 3.29 ( ⁄ ) + 0.72 + 408.59 log( ) 0.60 10/55 Drills = 10.75 log( ) − 961.32 0.74 5/33 Asphalt Distributors = 13.20 ( ) + 3704.04 ( 2 )⁄⁄ 0.63 12/332 Man Lifts = 8.82 log( ) + 45.51 0.95 7/67 Large Trucks with Special Body = 0.32 − 229.44 ( )2 + 1151.60 log( ) − 53.64 + 20.15 log( ) + 3696.48 5 + 3006.28 8 + 6582.92 11 0.71 45/3,585 designates the annual operating cost; all costs are in U.S. dollars. Equipment Type Operating Cost Estimation Model Adjusted R-Squared Sample Size After/Before Aggregation Dump Trucks = 12.3 × 10−9 )2(( )2 − 27.244 + 6262.58 ( 1 + 2 + 3 + 4 ) + 14161.67 5 + 13891.45 ( 6 + 7 ) 0.59 97/5,769 Pickup Trucks = 0.174 − 0.018 + 2719.20 ( 1 + 2 ) + 3442.98 ( 3 + 4 ) 0.63 72/8,721 Automobiles = 0.14 − 3.90 + 699.72 ( ) 0.5 − 0.73 ( )0.5 0.77 32/1,892 Vans = 0.36 − 3.85 + 145.84 0.69 49/1,903 Table 9. Equipment operating cost estimation models.

24 Table 8 also shows the adjusted R-squared values for each fitted estimation model; they ranged from 0.58 to 0.99. Several statistical tests (e.g., the Shapiro–Wilk test [12], Durbin–Watson statistic [13], and Breusch–Pagan test [14]) were also performed to confirm that the assumptions of the regression analyses were met; the results are shown in Table 10. The null hypothesis for the Shapiro–Wilk test determines if model residuals are normally distributed (as indicated by a high P-value). All equipment estimation models produced a high P-value except that for the mechanical street sweeper trucks indicating that the null hypothesis is not rejected and the residuals are distributed normally for these models. The null hypothesis for the Durbin–Watson test determines if the linear regression residuals are uncorrelated (as indi- cated by a high P-value). For a 95% confidence interval, the Durbin–Watson test did not reject the null hypothesis for any model, except the model for predicting the annual engine hours of drills, indicating that there is no autocorrelation between residuals in these models. The null hypothesis for the Breusch–Pagan test determines if the variance is constant for all residuals (as indicated by a high P-value). For a 95% confidence interval, all models provided homoscedastic residuals with a high P-value except for the model for estimating the annual mileage of man lifts, indicating that there was no heteroscedasticity in the residuals of these models. 3.8 Developing the Equipment Fleet Utilization Management Program Equipment utilization management plans can be influ- enced by several factors including budget, utilization thresh- old, and agency needs (i.e., demand levels for each equipment type). These constraints make fleet utilization management challenging, even when the number of equipment assets is not large. The research team developed a mathematical opti- mization program based on the proposed equipment utili- zation management framework; users only need to provide input data. The program incorporates statistical models that formulate the utilization cost functions and ensure that the needs for each piece of equipment are satisfied (Figure 5). Equipment Type Predicted Factor Sample Size Adjusted R-Squared Shapiro Test P-Value Durbin–Watson Test P-Value Breusch–Pagan Test P-Value Dump Trucks Annual Mileage 98 0.65 0.455 0.080 0.056 Pickup Trucks Annual Mileage 78 0.89 0.432 0.634 0.717 Automobiles Annual Mileage 32 0.74 0.419 0.636 0.568 Vans Annual Mileage 55 0.80 0.752 0.164 0.388 Sport Utility Vehicles Annual Mileage 37 0.81 0.947 0.064 0.367 Trailers Annual Mileage 72 0.68 0.759 0.080 0.944 Frequency of Usage 39 0.67 0.171 0.996 0.677 Front Loader Trucks Annual Mileage 17 0.83 0.644 0.952 0.291 Graders Annual Mileage 24 0.87 0.962 0.966 0.851 Annual Engine Hours 6 0.87 0.967 0.356 0.709 Mechanical/Air Street Sweeper Trucks Annual Mileage 29 0.92 0.525 0.338 0.911 Annual Engine Hours 19 0.75 0.558 0.142 0.518 Riding Mowers Annual Mileage 15 0.60 0.995 0.550 0.217 Frequency of Usage 12 0.89 0.837 0.814 0.944 Truck Tractors Annual Mileage 31 0.77 0.764 0.154 0.356 Snow Removal Attachments Annual Mileage 32 0.74 0.760 0.618 0.846 Rollers Annual Mileage 25 0.91 0.514 0.302 0.303 Annual Engine Hours 11 0.54 0.695 0.800 0.311 Frequency of Usage 18 0.74 0.627 0.436 0.748 Drills Annual Mileage 6 0.99 0.810 0.382 0.132 Annual Engine Hours 5 0.75 0.521 0.044 0.384 Asphalt Distributors Annual Mileage 13 0.75 0.678 0.326 0.237 Attachments Annual Mileage 18 0.75 0.862 0.180 0.072 Man Lifts Annual Mileage 18 0.76 0.821 0.248 0.042 Annual Engine Hours 6 0.59 0.408 0.700 0.446 Large Trucks with Special Body Annual Mileage 76 0.86 0.367 0.744 0.846 Table 10. Model validation results.

25 The optimization framework aims to minimize the total cost and improve equipment performance (through more efficient utilization management) and simultaneously sug- gests utilization values for each equipment type and fleet composition in each region to meet the demand. The opti- mization framework provides information on how the total operating cost of equipment is influenced by different factors (e.g., age, fleet size, and annual maintenance cost). The framework also compares the observed utilization of each equipment type with the optimal values and provides recommendations for changing these variables (e.g., usage value, fleet size, and allocation of the fleet to different counties or regions) to balance utilization. For example, the framework determines how an agency can reach the optimal utilization level for dump trucks by changing usage strategies (e.g., by relocating a portion of the trucks to other regions in the state). 3.8.1 Model Formulation This section introduces a mathematical optimization program that incorporates several models to estimate the optimal utilization values and fleet composition in each region in a state. The definitions of all sets, decision variables, and parameters used in these models are listed in Table 11. The objective of the optimization program is to deter- mine the lowest annual total costs (i.e., fixed and operating costs). The operating costs are influenced by fleet size in each region, which is highly dependent on the demand in the following year. Purchase, relocation, fuel, maintenance, and repair costs of equipment affect the required fleet size and complicate decision making. The objective function of the mathematical optimization program estimates the lowest total fleet management cost for specific region j in a specific year t. This cost includes the operating cost, equipment purchase cost, and equipment relocation cost and is presented by the following expression: ∑∑∑ ( ) + µ + β γ+ + + + + + + === − min , (1)1 1 1 1 1 1 1 110 1 c f u n lijt ijt ijt ijt ijt it jq ijqt q J j J t T where i denotes the equipment type and q denotes the region to which equipment is relocated. The first term in the expression is the cost of keeping an equipment unit in-service (estimated from the predictive annual operating cost functions presented in Table 9). The second term is the cost of adding new units of equipment type i to the fleet in region j in the following year, estimated by multiplying the purchase cost by the number of equip- ment units purchased. The third term is the cost of relocating Decision Variables Fleet size of equipment type ∈ in region ∈ at year ∈ Number of equipment type ∈ units purchased in region ∈ in year ∈ Number of equipment type ∈ units to be salvaged in region ∈ in year ∈ Number of equipment type ∈ units to be moved from region ∈ to region ∈ \{ } in year ∈ Average utilization level required to meet the demand for equipment type ∈ in region ∈ in year ∈ Input Parameters Cost per mile for moving one unit of equipment type ∈ in year ∈ The distance between region ∈ and region ∈ \{ } Total demand for equipment type ∈ in region ∈ in year ∈ Maximum allowed utilization level for equipment type ∈ Purchase cost of one unit of equipment type ∈ in region ∈ in year ∈ Total number of time periods Total number of counties/regions in a U.S. state Note: is the number of discrete time periods (i.e., fiscal years) in the planning horizon. Table 11. Definition of sets, decision variables, and parameters. Figure 5. Role of prediction models in the fleet utilization management framework.

26 equipment type i from region j to region q in the following year, estimated using the distances between counties/regions and the transportation cost. The objective function excludes salvage costs. The cost functions capture the effect of fleet size and utili- zation metrics (e.g., annual mileage) on operating costs for different inputs in each region (e.g., annual downtime hours, in-service age, equipment class). The annual operating cost is the summation of the annual fuel cost, unscheduled repair cost, and scheduled maintenance cost. Equation (1a) presents a general form of the operating cost function cijt+1 ( f ijt+1, uijt+1), which represents the annual operating cost of equipment type i in region j in year t + 1 (i.e., the following year) as a function of fleet size f ijt+1 and utilization value uijt+1. ( ) = α + w+ + + + +, (1a)1 1 1 1 1c f u f uijt ijt ijt ijt ijt Parameters α and w reflect the effects of the fleet size and utilization metric variations, respectively, on the annual operating cost and found in the model fitting process. To ensure that the total demand will be met for equipment type i in region j, the value of f ijt+1uijt+1 shall be equal to total demand. Therefore, the cost function (1a) can be re-formulated as: ( ) = π + α + w+ + + + + (1b)1 1 1 1 1 c f f D f ij t ij t i ij t ij t ij t The mathematical program imposes the following constraints: (2)1 1 1 1 1 1 1 ∑ ∑= + − − ++ + + + = + = f f n s l lijt ijt ijt ijt ijqt q J ijq t q J ∑ ≤+ = (3)1 1 l fijqt q J ij t ≥+ + (4)1 1 f D U ij t ij t i max ≥+ 0 (5)1nijt ≥+ 0 (6)1sijt ≥+ 0 (7)1lijqt Several factors such as the current fleet size, number of pur- chased and salvaged fleet, and number of the relocated fleet between counties/regions affect the decision on next year’s fleet size; these are considered in the optimization model. Constraint (2) ensures consideration of the decision on the fleet size in the following year. Fleet size of equipment type i will be updated by (a) adding the number of purchased equipment nijt+1 and “brought-in” equipment from other counties/regions into region j ∈ J, i.e., Σ Jq=1lijqt+1 and (b) sub- tracting the number of salvaged equipment sijt+1 and “sent-out” equipment from region j to other counties/regions Σ Jq=1lijqt+1. Constraint (3) ensures that the number of equipment units leaving region j to other counties/regions in the follow- ing year does not exceed the currently available fleet size in region j (i.e., a region cannot send more equipment units than it currently has to other counties/regions). Constraint (4) ensures that the demand will be met by having sufficient fleet size for the following year. The value of +1D U ij t i max provides the minimum required fleet size in region j for the following year to satisfy the demand. These constraints also ensure that the fleet is not driven more than the maximum allowed annual mileage per year. Constraints (5), (6), and (7) ensure that the number of equipment purchases, salvages, and relocations will get non-negative values in the following year. 3.9 Validating the Equipment Fleet Utilization Management Program Validation of the equipment fleet utilization management program is described in Section 2.9.

27 Summary and Suggested Research 4.1 Summary This research was conducted to develop a guide for utili- zation measurement and management of fleet equipment and a software package for use by state highway agencies. The research team reviewed the literature pertaining to equipment utilization measurement and management and performed an agency survey to identify the factors influenc- ing equipment utilization and management strategies. The research team then conducted further investigations and analysis and developed the guide and related software. Annual mileage, annual engine hours, usage over the last 12 months, and frequency of use were the utilization metrics most widely used by state DOTs. This research identified annual mileage as the recommended utilization metric for mobile equipment having an odometer, engine hours for stationary equipment with an engine, and frequency of use as the metric for other equipment. The research showed that the demand for equipment, in-service age, fleet size in a region, annual downtime hours, equipment class, utilization level in the previous year, and cost factors (including purchase, fuel, scheduled maintenance, and unscheduled repair costs) contribute to the utilization value of fleet equipment. In response to a survey of state DOTs, 19 state DOTs indicated availability of data for use in this research. After processing, cleaning, and reviewing the data received, the data from California, Louisiana, Michigan, New Hampshire, Pennsylvania, Utah, and Washington DOTs were selected for use in this research. A utilization management frame- work was proposed and utilization estimation models for 19 equipment types were developed. These equipment were dump trucks, pickup trucks, automobiles, vans, sport utility vehicles, trailers, front loader trucks, graders, mechanical street sweeper trucks, air street sweeper trucks, riding mowers, truck tractors, snow removal attachments, rollers, drills, asphalt distributors, attachments, man lifts, and large trucks with a special body. The data were aggregated at the region level over all years; each model estimated the annual mileage for a specific equipment type in a region. These models were developed through a rigorous model fitting process and considering various model structures for each equipment type (e.g., linear, quadratic, power, logarithmic, and nonlinear forms). A statistical correlation test was used to ensure that there was no multicollinearity between the independent variables; other statistical tests were used to confirm applicability of the assumptions of regression analysis. The research team then incorporated equipment utilization management into a mathematical optimization program. The program minimizes total cost by optimizing (a) fleet size, (b) number of equipment units to be purchased, (c) number of equipment units to be salvaged, (d) number of equipment units to be relocated to another region, and (e) average utili- zation level required to meet the demand for an equipment unit, for each year and region. These minimum costs were estimated for each equipment type separately and defined by the cost of keeping a piece of equipment in service, the cost of adding new equipment to the fleet in a region in the following year, and the cost of equipment relocation consid- ering the distance between counties/regions and the trans- portation cost rates. The program updates the fleet size in each region in the following year considering the number of purchased equip- ment units, “brought-in” or “sent-out” equipment from/to other regions, and salvaged equipment, and ensures that: 1. The number of equipment units leaving each region to other regions in the following year does not exceed the number currently available in that region; 2. The demand will be met by having sufficient fleet size in the following year (i.e., the product of fleet size and average utilization of each equipment type in each region in the following year shall be equal to or greater than the total demand); C H A P T E R 4

28 3. No equipment will be utilized in excess of its maximum allowed utilization level; and 4. The number of purchased, salvaged, and relocated equip- ment units will not be less than zero for the following year. The fleet size of an equipment type in each region is highly dependent on the demand in the following year as well as purchase, relocation, fuel, maintenance, and repair costs of the equipment. The fleet size in the next year depends on the current fleet size, number of purchased and salvaged equip- ment, and number of equipment units relocated between counties/regions. Therefore, the mathematical model con- siders the cost per mile for relocating equipment units of each equipment type in each year, the distance between each two counties/regions, the total demand for an equipment unit of a specific type in each region in each year, the maximum allowed utilization level for an equipment unit of a specific type, and purchase cost of an equipment unit of a specific type in each region in each year. The numerical results showed that by applying the equipment utilization management framework, the total costs (including capital investments and operating costs) are reduced. The research team developed a guide for utilization mea- surement and management of fleet equipment (presented in Part II). The guide integrates the utilization management framework and discusses the data input requirements and utilization measurement models; these models are embedded in the UPM software. Part III is the UPM’s user manual. 4.2 Suggested Research Several models were developed in this project based on data obtained from seven state DOTs. Improved models could be developed by using data from a larger number of agencies and considering factors such as seasonality and ownership. These improved models could then be used to upgrade the software developed in this research. Because of the complexity of the process for utilization prediction, training of fleet managers and other interested parties on the use of the software is suggested. Such train- ing would facilitate use of the software and achieve the desired utilization prediction. Developing training material and holding training sessions in the form of workshops, webinars, and other means would facilitate use of the devel- oped software.

29 References 1. Lauria, P. T. 2011. “Challenges and Opportunities: A Strategic Plan for Equipment Management Research.” NCHRP Project 20-7/ Task 309. Transportation Research Board of the National Acad- emies, Washington, D.C. 2. Laird, M. 2013. “How to Manage Fleet Data.” Wireless Machine- Information Systems. Available: https://www.construction equipment.com/how-manage-fleet-data (As of Sep. 10, 2019). 3. Osterstock, T., W. Kidd, A. Eliason, and P. Hicken. 2005. A Perfor­ mance Audit of the Division of Fleet Operations. Utah. Legislature. Office of the Legislative Auditor General. 4. BEB Industrial Asset Management. 2015. “Optimize your assets – fast and easy.” Available: http://www.bebsoft.com/index.html (As of Sep. 10, 2019). 5. Imhoff, C., and C. White. 2006. “Master Data Management: Creating a Single View of the Business.” Powell Media, LLC. Intelligent Solutions and BI Research. 6. Mercury Associates, Inc. 2005. NASA Official Fleet Management Handbook. National Aeronautics and Space Administration. Available: https://nodis3.gsfc.nasa.gov/npg_img/chapterview/ chapter-0preface.htm (As of Sep. 10, 2019). 7. Cook, R. D. 1977. “Detection of Influential Observation in Linear Regression.” Technometrics, Vol. 19, pp. 15–18. 8. Cook, R. D. 1979. “Influential Observations in Linear Regression.” Journal of the American Statistical Association, Vol. 74, pp. 169–174. 9. Efroymson, M. A. 1960. “Multiple Regression Analysis.” Mathemat­ ical Methods for Digital Computers, pp. 191–203. 10. Halinski, R. S., and L. S. Feldt. 1970. “The Selection of Variables in Multiple Regression Analysis.” Journal of Educational Measurement, Vol. 7, pp. 151–157. 11. R. Core Team. 2013. “R: A Language and Environment for Statistical Computing.” R Foundation for Statistical Computing. 12. Shapiro, S. S., and M. B. Wilk. 1965. “An Analysis of Variance Test for Normality (Complete Samples).” Biometrika, Vol. 52, pp. 591–611. 13. Durbin, J., and G. S. Watson. 1971. “Testing for Serial Correlation in Least Squares Regression. III.” Biometrika, Vol. 58, pp. 1–19. 14. Breusch, T. S., and A. R. Pagan. 1979. “A Simple Test for Heteros- cedasticity and Random Coefficient Variation.” Econometrica: Journal of the Econometric Society, pp. 1287–1294. 15. Yost, D. 2013. Ohio Department of Transportation Performance Audit. Ohio Department of Transportation. 16. Management Partners Inc. 2007. City of San Bernardino Fleet Services Assessment and Utilization Analysis. San Bernardino Fleet Services Division. 17. Minnesota Department of Transportation. 2012. Minnesota State­ wide Highway Systems Operation Plan. Minnesota Department of Transportation. 18. CTC & Associates LLC. 2012. Fleet Equipment Asset Management Performance Measures. Caltrans Division of Research and Innova- tion. Available: https://dot.ca.gov/-/media/dot-media/programs/ research-innovation-system-information/documents/preliminary- investigations/fleet-equipment-performance-measures-pi-a11y.pdf (As of Sep. 10, 2019). 19. Kauffmann, P., J. Hildreth, and R. Williams. 2013. Fleet Manage­ ment Performance Monitoring. North Carolina Department of Transportation. 20. Kauffmann, P., E. Howard, J. Yao, D. Harbinson, N. Brooks, R. Williams, and C. Gurganus. “Criteria for Fleet Management: Identification of Optimal Disposal Points with the Use of Equiva- lent Uniform Annual Cost.” Transportation Research Record: Journal of the Transportation Research Board, No. 2292, 2012, pp. 171–178. 21. Kader, M., T. L. Ramani, J. D. Johnson, C. Speigelman, J. Zietsman, and T. Jacobs. 2014. Fleet Equipment Performance Measurement Preventive Maintenance Model. Texas Transportation Institute. 22. Vance, R. J., M. S. Renz, M. A. Coccia, P. Croyle, and J. J. Karchner. 2014. Optimizing the Life Cycle of PennDOT Equipment. The Pennsylvania Department of Transportation. 23. Wyrick, D. A., and B. Storhaug. 2003. Benchmarking Fleet Manage­ ment. University of Minnesota: Center for Transportation Studies. 24. Lee, S., J. Turner, M. S. Daskin, T. Homem-de-Mello, and K. Smilowitz. 2012. “Improving Fleet Utilization for Carriers by Interval Scheduling.” European Journal of Operational Research, Vol. 218, pp. 261–269. 25. Regan, A., H. Mahmassani, and P. Jaillet. “Evaluation of Dynamic Fleet Management Systems: Simulation Framework.” Transporta­ tion Research Record, No. 1645, 1998, pp. 176–184. 26. Figliozzi, M., J. Boudart, and W. Feng. “Economic and Environ- mental Optimization of Vehicle Fleets: Impact of Policy, Market, Utilization, and Technological Factors.” Transportation Research Record: Journal of the Transportation Research Board, No. 2139, 2011, pp. 1–6. 27. Jin, D., and H. L. Kite-Powell. 2000. “Optimal Fleet Utilization and Replacement.” Transportation Research Part E: Logistics and Transportation Review, Vol. 36, pp. 3–20. 28. Chen, C., and J. Lin. 2006. “Making an Informed Vehicle Scrappage Decision.” Transport Reviews, Vol. 26, pp. 731–748. 29. Lin, J., C. Chen, and D. A. Niemeier. 2008. “An Analysis on Long Term Emission Benefits of a Government Vehicle Fleet

30 Replacement Plan in Northern Illinois.” Transportation, Vol. 35, pp. 219–235. 30. Johnson, P. D. 2002. Principles of Controlled Maintenance Manage­ ment. The Fairmont Press, Inc. 31. Lee, W.-S., A. Y. Huang, Y.-Y. Chang, and C.-M. Cheng. 2011. “Analysis of Decision Making Factors for Equity Investment by DEMATEL and Analytic Network Process.” Expert Systems with Applications, Vol. 38, pp. 8375–8383. 32. Vujanović, D., V. Momčilović, N. Bojović, and V. Papić. 2012. “Evaluation of Vehicle Fleet Maintenance Management Indicators by Application of DEMATEL and ANP.” Expert Systems with Appli­ cations, Vol. 39, pp. 10552–10563. 33. Said, H., T. Nicoletti, and P. Perez-Hernandez. 2014. “Utilizing Telematics Data to Support Effective Equipment Fleet-Manage- ment Decisions: Utilization Rate and Hazard Functions.” Journal of Computing in Civil Engineering, Vol. 30, p. 4014122. 34. Roy, B., and P. Windover. 2015. Idle Reduction Assessment for the New York State Department of Transportation Region 4 Fleet. New York State Department of Transportation. 35. Kargul, A., Glaese, A., Kessler, S. and Günthner, W.A. 2016. “Heavy Equipment Demand Prediction with Support Vector Machine Regression Towards a Strategic Equipment Management.” Inter­ national Journal of Structural and Civil Engineering Research, Vol. 6, No. 2, pp.137–143.

Next: Part II - Guide for Utilization Measurement and Management of Fleet Equipment »
Utilization Measurement and Management of Fleet Equipment Get This Book
×
MyNAP members save 10% online.
Login or Register to save!
Download Free PDF

State highway agency equipment fleet assets are vital to the delivery of agency programs, projects, and services. Measuring, monitoring, and reporting on asset utilization levels are necessary for the management of the equipment fleet and meeting the highway agency’s business needs.

The TRB National Cooperative Highway Research Program'sNCHRP Research Report 957: Utilization Measurement and Management of Fleet Equipment is both a handbook on equipment utilization concepts and a guide for making cost-effective equipment utilization decisions.

The Utilization Prediction and Management Software allows the user to:

• estimate equipment utilization and manage the fleet at a region-level based on available measurable information

• identify equipment that is under- or over-utilized, needs to be salvaged, or needs to be relocated; and

• identify the fleet management strategy that minimizes the total fleet management costs.

  1. ×

    Welcome to OpenBook!

    You're looking at OpenBook, NAP.edu's online reading room since 1999. Based on feedback from you, our users, we've made some improvements that make it easier than ever to read thousands of publications on our website.

    Do you want to take a quick tour of the OpenBook's features?

    No Thanks Take a Tour »
  2. ×

    Show this book's table of contents, where you can jump to any chapter by name.

    « Back Next »
  3. ×

    ...or use these buttons to go back to the previous chapter or skip to the next one.

    « Back Next »
  4. ×

    Jump up to the previous page or down to the next one. Also, you can type in a page number and press Enter to go directly to that page in the book.

    « Back Next »
  5. ×

    To search the entire text of this book, type in your search term here and press Enter.

    « Back Next »
  6. ×

    Share a link to this book page on your preferred social network or via email.

    « Back Next »
  7. ×

    View our suggested citation for this chapter.

    « Back Next »
  8. ×

    Ready to take your reading offline? Click here to buy this book in print or download it as a free PDF, if available.

    « Back Next »
Stay Connected!