Click for next page ( 208


The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 207
Level . Basic Runcutting Chapter 5. Runcutting runcutting 5.1 Basic Runcutting The process of converting (or Runcutting is the process of turning blocks into work pieces or "runs" for drivers. cutting) vehicle blocks into work LEVEL The word cutting is used to describe the process of "cutting" blocks into pieces 1 assignments for operators. that drivers can actually work. A "run" may consist of all or part of a vehicle block, and may have single or multiple pieces. During the runcutting process blocks may run remain intact, be cut, or in some cases be completely rehooked. A work assignment for an operator. In practice runcutting is an extremely complex process that takes into account a range of fac- Most often, run refers to a whole tors, both qualitative and quantitative, to produce a set of outcomes that significantly affect day's work assignment. transit operations and an agency's budget. The scheduler is actually creating the work days for the bulk of the agency employees, significantly affecting the cost of operation. The sched- block ules department is in many ways a transit agency's budget implementation and enforcement A vehicle (or train) assignment that department. includes the series of trips operated We should note up front that runcutting can range from simple solutions to infinitely complex by each vehicle from the time it pulls out to the time it pulls in. A problems. In fact, most computerized runcut "optimizers" are not true optimizers at all--they complete block includes a pull-out cut down the problem into something manageable and then optimize that solution. This shows trip from the garage followed by one just how complex the mathematics of runcutting can be--that even the most powerful com- or (usually) more revenue trips and puters are not able to mathematically solve complex runcutting problems. concluding with a pull-in trip back to the garage. Runcutting Objectives What are we trying to achieve when creating runcut solutions? There can be many objectives, and defining the objectives can be an extremely difficult task, raising more questions than it rehooking answers. But knowing the priorities at your agency is critical because without objectives one The process of changing how trips cannot measure the success of a runcut. Below we note some of the typical objectives a sched- are linked into a block. This is done uler will be trying to achieve when undertaking runcutting. when evaluating blocks and during the runcutting process. 1. "Legal" Solutions. First and foremost the runcut must be "legal." This is an expression used at many systems and because labor contracts are often involved, indeed there are times when it carries the force of law. "Legal" means that it must meet all the writ- ten rules of the labor agreement. It must also meet any safety, driving time, or break The word "run" is often Tip requirements, some or all of which may be legislated. misused. Individual trips, nonstop portions of routes are often 2. Efficient Solutions. In most cases the scheduler will be trying to achieve an efficient called "runs" by laymen. Do not stray runcutting solution. In bus transit agencies the operator cost typically accounts for from the appropriate definition and more than % of total operating cost. Therefore the impact of runcutting efficiency use. is significant in overall operating efficiencies. Exactly what defines "efficiency" can at 5-3

OCR for page 207
Chapter 5. Runcutting Level . Basic Runcutting straight run times be somewhat difficult to determine and will be discussed in some detail in the fol- A run in which trips are consecutive lowing sections. without interruption. Straight runs 3. "Streetable" Solutions. The scheduler should create runcuts that maximize the op- do not contain any breaks (except erational potential of the solution. For the same cost and legality, comparative runcuts for meal breaks at some systems) could diminish (or alternately enhance) the ease of operation. For example, in systems for the operator. Any break in a which have relatively high "leave" benefits, i.e., vacation days, personal days, sick days, straight run is usually less than one birthdays off, etc., a streetable solution might mean more runs markedly longer than hour in length. A straight run with a hours, often in the - to -hour range ( -day work week). break is referred to as a multipiece 4. "Friendly" Solutions. In addition to the above objectives the scheduler should be trying straight. to achieve results that allow for a "reasonable" workplace for the operators. Schedulers need to be aware that the runs and rosters developed define an operator's entire work split run life, and should look to achieve a balance of operator requests and preferences where A run containing two or more pieces possible. of work separated by a break over one hour in length. Also known Of course some of the above objectives are at times in conflict or at the very least difficult to as a swing run. At some systems, achieve simultaneously. That is where the skill of the scheduler, able to balance the varying three-piece split runs are allowed, objectives, comes into play. but one of the breaks (or "swings") is usually paid whereas in two-piece Types of Runs split runs the break is generally not Runcutting, as described above, is creation of operator assignments or runs. But what are paid. Split runs tend to be used these runs? The types of runs operated by public transit agencies generally fall into three to allow both peaks to be covered broad categories. These are described below. by one operator since the work day would otherwise be too long for a 1. Straight Runs. This type of run implies work paid as "straight through" or continuous straight run. time on duty. Traditionally a straight run has consisted of a single piece of work, where the operator stays on the same vehicle for the whole day. A second type of straight run swing time involves a break (often required by labor agreement or legislation) between two pieces of work. This break may be paid or unpaid, and be taken on the vehicle or at the depot, The elapsed time (usually unpaid) according to site-specific labor agreement provisions. The break in a straight run is between the pieces of a split run. often between and minutes in duration. Also known as "intervening time." if swing time is paid, it is sometimes called "inside time." Below are graphical examples of straight runs. spread time Total time between the start of the first piece and the end of the last 1 In some systems with both types of these straight runs, only the former is called a "straight," the latter, even with a break under 60 minutes, is considered at that system to be a split (or swing). For simplicity, this document distinguishes them as one-piece straights piece of a split run with two or more and multipiece straights. pieces. Also known as "outside 5-4 time."

OCR for page 207
Level . Basic Runcutting Chapter 5. Runcutting tripper A short piece of work whose total 2. Split Runs. Split (or Swing) runs refer to runs that have two pieces, with a (usually) time is less than that specified as longer unpaid break between those pieces. The operator is not on duty between the constituting a full-time run. A tripper pieces of work, and typically all pieces start and finish at the home garage. The break in is often a piece of work in the AM a split run is characteristically longer than for straight runs (both kinds), greater than or PM peak period that cannot be minutes, and may be as long as three or four hours. Split runs tend to be used to allow combined with another piece of both peaks to be covered by one operator since the work day would otherwise be too work to form a split run because of long for a straight run. insufficient hours, excessive swing time, or excessive spread time. Below is a graphical example of a typical split run. Trippers are often operated by extraboard or part-time operators. Tripper can also refer to a vehicle that pulls out, makes no more than one round-trip, and pulls in. 3. Trippers. Trippers are almost always short one-piece straight runs, and are often used in peak periods. Trippers are sometimes known as "part-time runs." There are many different Tip issues to juggle when developing split runs. Before you begin, know whether your property has rules regarding the total length In some cases the difference between types of runs can become fuzzy--particularly of the driving day including the between a split run and a straight run with an unpaid break, or between a tripper with unpaid break times, the length of multiple pieces and a split run. Invariably, though, the unique conditions identifying time a driver can drive, whether runs each run type can be found in your labor agreement (and it is likely that your agreement with more than two work pieces are clarifies the differences between very similar runs). allowed, i.e., three-piece runs, and the trade-offs between split runs and These examples are basic in nature and do not cover all types of runs or all types of applications trippers. of those runs. For example, a split run does not necessarily cover both peaks, and in fact may be used on weekends or where there is little peak effect. Components of Runs Runs consist of a number of distinct components--the actual in-vehicle time is only one of them. Importantly, the actual driving time is only one of several components and may account for as little as % or % of the total paid time, depending upon the type of run and labor agreement provisions. 5-5

OCR for page 207
Chapter 5. Runcutting Level . Basic Runcutting The diagram below provides a graphic visualization of the possible components of a run. It is work hours interesting to note than in the split run shown the actual driving time is responsible for only Total hours worked by an operator, two of eleven unique run components. not including fringe benefit hours such as sick leave, holiday, etc. Work hours include only labor hours associated with the requirements of putting the runs in service and operating the service. collaterals All of the various types of penalties and premiums that might be required to make legal runs. premium pay Total work hours and spread hours are indicated as these are numbers that are needed to Pay to an operator that is over and calculate penalties such as overtime, spread, and shift premiums. These penalties or premi- above the straight time pay rate; ums--which in some systems are collectively known as "collaterals"--are discussed in more includes overtime premium, spread detail in later sections. At this stage it is sufficient to note that understanding the components premium, shift premium, and any of runs is required to develop an understanding of the runcutting process. other operating premiums as defined by the contract. Runcutting Outputs The outputs of any scheduling process can be broken down into two categories--the actual operational outcomes, and the reports required describing those outcomes. In the case of runcutting the outcomes are simply a set of operator assignments that are valid, cover all of the blocks, and meet the objectives of the agency. However there are work units, i.e., departments, divisions, offices, etc., of an organization that require a description or sum- mary of those runs--in effect, the tools with which certain people will be able to undertake their roles. Some of these are listed in the table below. 5-6

OCR for page 207
Level . Basic Runcutting Chapter 5. Runcutting overtime premium Output Description Used By Pay at the rate of 1.5 times (or higher) the normal rate for work Run Guide A summary of runs that Schedulers, garage staff performed in excess of daily or describes start/finish weekly thresholds, usually eight or locations, work hours, and ten hours per day or 40 hours per cost element breakdowns week. Runs Summary List of runs showing start/ Payroll systems finish times, hours worked, and paid hours shift premium Dispatch Sheet A list of all runs, sorted by Garage staff, to track staff and vehicle A premium paid to operators for start time movements throughout the day working during times of the day that are subject to special pay differen- Runcut Statistics Summary of costs, totals, Schedulers for summaries/ tials, e.g., an owl (late night/early penalties, etc. comparisons, Senior Management for morning) run. costing analysis The Role of Computers spread premium You will recall the comment that many runcutting problems are qualitative in nature and are, Pay equal to one-half or more of therefore, too complex for any computer, alone, to optimize completely. This does not, how- all minutes in excess of a specified ever, mean that computers do not have a significant role to play as tools to significantly assist maximum spread time, in addition in developing runcuts. to regular straight pay. The spread premium may be multilayered, e.g., Computerized Scheduling Packages half of all time up to 60 minutes over the specified maximum spread The advent of computerized scheduling packages containing complex algorithms has enabled time and three-quarters or all time significant improvements in the ability to generate efficient runcut solutions. However, since more than 60 minutes over the the computer is trying to solve what it sees as a mathematical or quantitative problem, the specified maximum spread time. inputs need to be numeric. So as a scheduler you will need to resolve issues of preference in Spread premium is separate and terms of weights or penalties. And this requires skill and practice. For example, do you think distinct from overtime premium. that longer spreads are less preferred than having additional trippers? If so, by how much-- %, %? The automated algorithms require the user to be able to define such preferences concisely and numerically. The modern ability to run numerous runcut solutions and compare results is a significant ben- efit generated by automated runcutting programs over previous manual-only runcuts. A key point to remember is that the computer is only a tool to be used by the skilled scheduler, to produce high-quality solutions. Keeping this in mind, here are a few considerations when using computerized systems to generate runcuts: Never accept the automated solution as given. Always consider the need to fine tune--for operational preference, runcut desirability, or whatever reason. Very rarely 5-7

OCR for page 207
Chapter 5. Runcutting Level . Basic Runcutting is an automated solution ready for implementation without the need for some, even if minor, manual manipulation. Run multiple solutions, each time changing (preferably) one constraint at a time. The direct impact of this particular constraint can then be measured. Then do the same report allowance process with constraints that impact on each other. This way, the best set of inputs can The amount of time paid an opera- be developed. tor from sign-in time to pull-out time. Experiment, review, change, and adjust. Use the interactive tools to review and ad- During this time, the operator may obtain instructions and supplies just solutions. This may lead you to adjusting parameters, weights, costs or constraints, pertinent to his/her run, locate the and resubmitting for a complete new (and improved) solution. assigned vehicle, and perform a Use the tools provided by the system to work interactively. pre-trip inspection. Use the tools provided by the system to double check accuracy of a runcut solution. Familiarize yourself with all of the ways the system can summarize, present, and al- sign-on time low you to review information. (sometimes called sign-in time) The time an operator is assigned to Our experience is that users tend to focus on a subset of the available functionality of comput- report for duty at the start of each erized systems and as a result do not exploit the full capabilities of the software. This in turn piece of a run. The operator may affects either the efficiency or quality of runcuts created. be required to sign in or may be acknowledged by the dispatcher as Basic Spreadsheet Tools having reported. The spreadsheet is not forgotten as a means of developing runcuts either. Spreadsheets pro- vide a range of capabilities. Some typical examples include: clear allowance Calculation. Use formulas as much as possible to calculate totals, penalties, costs and The amount of time paid to an operator at the conclusion of the to generate summaries. For example, if the report allowance is minutes, and you run to turn in transfers, fare media, want to show both garage depart and sign-on times, the sign-on time field should or other supplies and reports. equal the garage depart time less minutes (expressed in a formula as . x or as time(: ,: ,: )). If the clear allowance is minutes, then the sign-off time would equal the garage arrival time plus minutes. There are endless examples, at varying levels of sophistication, where calculations can be automatically generated. In addition, times can be converted from time to decimal format if so desired. If the entry in cell A is : , the formula is: =hour(a )+minute(a )/ , which yields . . Learn the things your Visualization. Use formatting to show a variety of information. For example, use Tip conditional formatting to show if something is not legal (e.g., if the total spread field runcut program does best is greater than hours and the maximum allowed is hours, highlight the cell with and use it to your full advantage. Not a red background). Conditional formatting in spreadsheets is a powerful tool that can all programs are equally good at all aspects of runcutting. greatly assist the scheduler. Validation. Use formulas to check whether rules and preferences are being broken. These can be applied as checked fields, or as formatting changes, highlighting issues. 5-8

OCR for page 207
Level . Basic Runcutting Chapter 5. Runcutting Finally, and this cannot be stressed highly enough, do not type values into cells unless ab- solutely necessary! The more numbers that are typed, the greater the chance for error. Use formulas, calculations, and formatting as much as possible. Learn to use lookup tables and pay-to-platform ratio other functions that can assist to streamline the runcutting process, and can reduce errors. As The ratio of pay hours to platform with computerized scheduling packages there is a tendency for schedulers to use only a small time. For example, if an operator portion of the capacity and power of spreadsheet systems. receives 9:00 in pay for 8:00 of platform time, the pay-to-platform How to Measure the Success of a Runcut ratio is 1.125 (9:00/8:00). The pay- The success of a runcut is always measured against the standards and objectives of the agency. to-platform ratio is one of the most widely used methods of measuring Usually based on agency historic data as a foundation, pay-to-platform ratios are the pri- runcut efficiency and is often used mary guide (more about pay-to-platform later in this section). Some typical objectives were to measure the impacts of non- discussed above. As with many aspects of scheduling, be thorough and methodical. Develop a platform items (such as report or template for comparing one solution to the next (spreadsheet is preferred), each with a score travel times) on operator pay hours. or summary for the key criteria. Only when the relevant information is laid out can an objective Some systems use the inverse, the assessment then be made. ratio of platform to pay hours. Remember, as a scheduler you will be asked to provide advice regarding the differences or im- pay hours pacts from one runcut to the next. And even before you reach this point you are likely to com- pare your initial or preliminary solutions as you consider alternative solutions. The number of hours for which an operator is paid at his/her rate. Pay The example template below compares some key cost measures between two runcuts. A de- hours include work hours, make-up bate about the merits of one solution against another will be put off until later in this chapter. time, overtime premium, spread Suffice to say, the tables allow us to see quickly that Solution uses more part-time operators premium, and any other adjust- ments called for in the contract. but achieves a lower cost. The agency objectives and experience can then be used to determine whether that is a desirable outcome or not. Linking spreadsheets Tip with "look-up" formulas minimizes direct entry and allows you to keep all of your input variables on a single master sheet. 5-9

OCR for page 207
Chapter 5. Runcutting Level . Basic Runcutting Runcut Comparison Solution 1 Solution 2 Straight Runs 40 73% 38 64% Before beginning the Split Runs 12 22% 9 15% Tip Total FT Runs 52 95% 47 80% runcut, take a broad view at your blocks. Understand whether Part Time Runs 3 5% 12 20% the same level of service operates all day, which lends itself to straight Total Operators Required 55 59 runs, or whether service is "peaked" with more vehicles (and drivers) Hours Breakdown Total Avg % Total Avg % required during limited times of day. Revenue 420.0 7.6 88% 420.0 7.1 88% This lends itself to a combination of Report 22.3 0.4 5% 22.7 0.4 5% part-time and split runs in addition Travel 33.5 0.6 7% 34.0 0.6 7% Total Work Hours 475.8 8.7 476.7 8.1 to a full-time and straight runs. The more variation in block sizes, the Penalties more likely you will have a complex Spread 14.4 10.8 runcut. Overtime 17.9 2.3 Guarantee 6.1 2.3 Total Paid Hours 514.3 492.1 Pay/Plat Ratio 1.2244 1.1717 Note that total operators could potentially be shown in terms of Full-Time Equivalent (FTE) operators. In such cases part-time operators may be substituted on a for basis. For Solution the FTE would be . , and for Solution the FTE would be , using this approach. Be care- ful, though, since that approach works well where part time operators are covering the peaks and have shorter average hours, but may be misleading if the part-time runs are longer or not only covering the peaks. The approach to definition of FTE, and the applicability of FTE as a measure to compare runcuts, will vary by agency. Some agencies just view each employee as part of a headcount, and some use an FTE approach. Runcutting Inputs The success of any runcut depends on the quality of the information available as inputs to the runcutting process. This applies as much to manual runcutting as it does to computerized run- cutting. If the basic elements of a runcut are not well defined, the result will not be effective. 5-10

OCR for page 207
Level . Basic Runcutting Chapter 5. Runcutting Scheduling requires absolute accuracy of details. To this end, make sure you have a good meth- DO NOT attempt a runcut odology for ensuring that when you start to runcut, whatever the size of the problem, you have Tip the following pieces of information, in their entirety: until you know your blocks are complete and correct A complete set of trips and vehicle blocks; and all work rule preferences are All relevant defined rules (usually the labor agreement summarized); defined. Defined relief types, relief locations, and travel times; and Known limitations--cost limits, work rule preferences, etc. Most automated scheduling packages have a means for checking that all the required informa- travel time tion is available to start runcutting. Always double check, though. You can simply go back and Paid time allowed for an operator check your vehicle stats against what they should be. For spreadsheet-based methods, make to travel between the garage and sure you have a checklist, and a method for viewing both the runs and blocks details. Checklists a relief location. If the travel is for are important throughout the scheduling process, to ensure work is undertaken thoroughly, in relief purposes only and is not part a methodical manner. of a pull-in or pull-out, then travel time is not included in platform time. Trips/Blocks First and foremost we need something to runcut. The runcut process starts with a complete service curve and correct set of vehicle blocks and trips. Where blocks define the assignment for a given vehicle, the run will define the assignment of a given driver. A plot of the number of buses in service by hour. We can, at this stage, consider the nature of the blocks and runs we are working with. In par- ticular the nature of the "service curve" can affect how we will approach the runcutting pro- cess. For example, is the service highly "peaked" (i.e., has a lot of blocks with peak-only trips), or do the same frequencies operate pretty much all day? The answer to this question will have a significant bearing on how the runcut results will look and on how we approach the runcut. The Broad Street example will be used to show how to lay this out visually. Remember that while scheduling is a very detailed, numbers-based process, the best way to understand the data is often by looking at things visually, usually with graphs. This allows a simple representa- tion of a great deal of detail that would otherwise be difficult to understand. Below is a simple "Vehicles in Operation" graph. This graph simply shows how many buses are in operation anywhere at any given time during the day on Route Broad Street. The graph can be generated quite simply in a spreadsheet. For computerized scheduling packages the graph can be generated automatically--so check out how to do so as this is an invaluable tool. 5-11

OCR for page 207
Chapter 5. Runcutting Level . Basic Runcutting vehicles in operation graph A graphical representation of the number of vehicles in operation by time of day, typically by route but also by garage or system. Looking at this graph, we can immediately see that there will be some complexity in cutting the runs during peak times. We can also assume that we may need to use different run types to cover the peaks, including (potentially) split runs or trippers. Some other things we can immediately note in the graph above include: Many of the blocks are around hours which will probably allow us to cut them into two pieces, each as a single run. The peak blocks are shorter in the AM Peak than the PM Peak (this is not unusual). The AM blocks are around : and the PM blocks around : . This may allow us to create : split runs without too much difficulty (subject to spread limitations). We will need two additional runs during the peaks to cover the two additional peak vehicles. We will have three runs in operation during the off peak. 5-12

OCR for page 207
Level . Basic Runcutting Chapter 5. Runcutting So by simple visual representation, without looking at any tables of numbers, we can discern a great deal of information about the vehicle blocks, their impacts on the runcutting to follow, and some potential outcomes. Of course not all graphs will have this form--some have unbal- anced peaks, some are flat, or have higher peaks, or wider/shorter spans. But in each case one can immediately consider implications for the runcutting process. Naturally, as we get into the actual cutting of blocks we will need to work at a more detailed level. Rules, Constraints, and Practices We now have a set of blocks to generate runs from. What else do we need? One thing you will hear about often when discussing runcutting is the word "constraints." In the context of creat- ing a runcut there are numerous constraints, each limiting how we can cut the blocks into runs. Constraints include the following types of categories: Hard Rules. "Hard" rules are typically defined in the labor contract. They include spe- cific fixed limits such as how many hours can be worked in a day, or how many of a type of run is allowed. Sometimes hard rules are expressed as a fixed number (e.g., a -hour maximum work time in a day). In some cases they are expressed as percentages (e.g., no less than % of full-time runs may be straight runs). At times the language and definition of hard rules can be extremely complex. In this context hard rules cannot be broken and are nonnegotiable. Soft Rules. "Soft" rules may be in the labor contract or may fall into the category of "operating practices" (see below). Soft rules tend to be preferences or "wants," rather than nonnegotiable requirements. Typical soft rules tend to relate to not scheduling excessively to hard rule requirements. For example, while you may be allowed to sched- ule up to a -hour spread, there is a preference to keep as many as possible within hours. This may be written into the labor agreement or may just be an understanding with operators. Operating Practices. Some constraints, both hard and soft, are not written into the labor agreement, or even necessarily required by operators, but are applied based on either historical precedents or by preference. For example, it may be an agency pref- erence that any operator who works on a certain route always gets a break every two hours (if that is a busy route). Or perhaps, going back to the spread example, the labor agreement allows -hour spreads but Management sees that as excessive and prefers to schedule to a -hour maximum. 5-13

OCR for page 207
Level . Basic Runcutting Chapter 5. Runcutting What else can our blocks tell us about the optimal runcutting solution? Let's go back and de- construct the "Vehicles in Operation Graph" from the previous section. What does it tell us? We can see there are three vehicles operating during the day and five during the peaks. Immediately that suggests we will have either two split or four trippers, to cover those addi- tional peak blocks. That leaves us with three all-day blocks, each operating well in excess of the allowable platform time for a straight run. These blocks will therefore need to be cut, potential- ly into two straight runs each. But as these blocks are only between and hours in length we can also expect shorter straight runs, possibly with some guarantee time. By this stage of our analysis we now have an idea of not only how many runs but also of the likely breakdown of those runs by run type and the potential for guarantee time--all without doing any runcutting at all! This is what we mean by "knowing the answer" before starting. 5-27

OCR for page 207
Chapter 5. Runcutting Level . Basic Runcutting As we've stated elsewhere, many scheduling tasks are made easier by simple graphical repre- sentation of the information rather than detailed numerical tables. This applies equally to run- cutting tasks. A good visualization tool is to lay out the blocks in a graphical manner as shown below. 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 Block 1 2 3 4 5 6 7 While the times are rounded to -minute segments, it enables us to pretty quickly see block lengths and suitability for run types. For example blocks one, four, six, and seven appear suit- able for split run or tripper pieces, and blocks two, three, and five appear suitable for straight runs. Creating the Runs The hard work is far from done. Although we have an idea of how the runcut will look, we need to work through the detail of that runcut and create the runs. Looking back at our graphical blocks diagram we have a good idea that blocks one and four will combine with blocks six and seven to create two split runs. And we expect to cut blocks two, three, and five into two straight runs each. However the graphic does not provide enough detail about specific times and locations. So we turn back to the schedule. Let's start by trying to cut Block into straight runs. Block starts at : and ends at : , totaling : . Our run gets minutes report allowance and so signs on at : . We'd like to be as close to eight hours as possible (to avoid guarantee or overtime) and so ideally we would like to finish at around : . Allowing for minutes of clear allowance, and minutes of travel time, that means we would like to cut the run at around : . But the only relief opportunities (i.e., the times that the bus passes Point A) in this time vicinity are at : or : . So we have a judgment call to make, and it has implications for not only the AM run but of course the PM run also. The best thing is to look at the two options and see the implications as below. 5-28

OCR for page 207
Level . Basic Runcutting Chapter 5. Runcutting Option 1 - Cut at 12:09 Start End Run Block Report Sign Off Plat Report Travel Total Spread Paid Pay/ Type Spread Guarantee Overtime # # Time Time Hours Hours Hours Hours Penalty Hours Plat Time Place Type Time Place Type 1 Str 2 5:46 6:01 Garage Pull 12:09 A Street 12:34 6:08 0:30 0:10 6:48 6:48 1:12 0:00 0:00 8:00 1.304 2 Str 2 11:44 12:09 A Street 18:38 Garage Pull 18:53 6:29 0:30 0:10 7:09 7:09 0:51 0:00 0:00 8:00 1.234 12:37 1:00 0:20 13:57 2:03 0:00 0:00 16:00 1.268 Runcutting, scheduling, Tip and planning are often Option 2 - Cut at 13:39 iterative processes. When you begin your runcut, you may notice Run Block Report Start End Sign Off Plat Report Travel Total Spread Paid Pay/ that adjusting the schedule slightly Type Spread Guarantee Overtime # # Time Time Hours Hours Hours Hours Penalty Hours Plat Time Place Type Time Place Type can either add service at no cost, 1 Str 2 5:46 6:01 Garage Pull 13:39 A Street 14:04 7:38 0:30 0:10 8:18 8:18 0:00 0:09 0:00 8:27 1.107 or could save significantly with a 2 Str 2 13:14 13:39 A Street 18:38 Garage Pull 18:53 4:59 0:30 0:10 5:39 5:39 2:21 0:00 0:00 8:00 1.605 slight trim. Good schedulers do not 12:37 1:00 0:20 13:57 2:21 0:09 0:00 16:27 1.304 just cut what's in front of them, but make sure everyone responsible The answer in both cases is "not good!" Both options result in substantial guarantee time. understands the implications of the Why? Block is only : , which allows an average of : of platform time. Even adding the schedule. travel, report, and clear allowances this only gets us to an average run length of : , mean- ing somewhere around an hour of guarantee time for each run. In this case we choose to cut the blocks at : since this incurs the minimum total cost. However we would be looking for other solutions as we cut the runs since we have incurred some serious inefficiency in the first two runs. In the end we will choose Option based upon the fact that it incurs less minutes of paid time. Is minutes important? The answer is an emphatic "yes!" In scheduling every minute matters. To illustrate this point let's consider the costing impacts of that minutes. Over a year (as- suming weekdays) that comes to a total of . hours. At a base hourly rate of that comes to over , , which will end up more once variable benefits are applied. From this example we can use a rule of thumb that every hour of paid time equates to over , annu- ally--that's why not wasting any paid time in every situation is critical. We will push ahead and cut straight runs from Blocks and (the other two all day blocks) in the same manner as we cut Block . Again we need to decide when to cut the block to mini- mize the overall paid time while keeping the runs legal. In this case if we cut Block at : the first run becomes less than six hours, so we are forced to cut it at : . We cut Block at : as that option best balances the work hours of the AM and PM runs. Our runs summary now looks like the following: 5-29

OCR for page 207
Chapter 5. Runcutting Level . Basic Runcutting Option 1 - All Long Blocks Now Cut Start End Run Block Report Sign Off Plat Total Spread Paid Pay/ Type Spread Guarantee Overtime # # Time Time Hours Hours Penalty Hours Plat Time Place Type Time Place Type 1 Str 2 5:46 6:01 Garage Pull 12:09 A Street 12:34 6:08 6:48 6:48 1:12 0:00 0:00 8:00 1.304 2 Str 2 11:44 12:09 A Street 18:38 Garage Pull 18:53 6:29 7:09 7:09 0:51 0:00 0:00 8:00 1.234 3 Str 3 5:35 5:50 Garage Pull 12:39 A Street 13:04 6:49 7:29 7:29 0:31 0:00 0:00 8:00 1.174 4 Str 3 12:14 12:39 A Street 18:19 Garage Pull 18:34 5:40 6:20 6:20 1:40 0:00 0:00 8:00 1.412 5 Str 5 6:05 6:20 Garage Pull 13:09 A Street 13:34 6:49 7:29 7:29 0:31 0:00 0:00 8:00 1.174 6 Str 5 12:44 13:09 A Street 19:53 Garage Pull 20:08 6:44 7:24 7:24 0:36 0:00 0:00 8:00 1.188 38:39 42:39 5:21 0:00 0:00 48:00 1.242 Again we can see that the runs are generally inefficient, each resulting in guarantee time. At about this stage we may want to go back, rethink the schedules, and note that it would be possible to extend the hours of operation, by one round trip for each vehicle, with minimal cost impact. The cost impact would be minimal as the additional hours would be absorbed by the available guarantee time, meaning the additional paid hours would be negligible. The real cost would be the variable vehicle costs, and these tend to form only a small percentage of total operating costs. But we should leave such decisions until we have worked through all the runs, both straights and splits. Next we might create split runs from the smaller peak blocks. We have four peak blocks (Blocks , , , and ), two in each peak. These should be able to form two split runs. We can simply list the appropriate blocks as follows to get an idea of how the runs will look. Block Garage Depart Garage Arrive Hours 1 5:46 9:19 3:33 6 15:01 19:49 4:48 Spread 14:33 Work Hours 9:11 4 6:05 9:38 3:33 7 15:20 19:19 3:59 Spread 13:44 Work Hours 8:22 5-30

OCR for page 207
Level . Basic Runcutting Chapter 5. Runcutting Now we have two split runs. The block lengths result in these runs being longer than the straights. We'll assume that is acceptable and push ahead. This gives us two splits and six straights, a total of eight runs. All blocks are covered and we must be done, right? Wrong! Go back and look at the split runs--they fail the very first test of meeting our scheduling criteria--they are not legal! The two split runs are well in excess of the -hour maximum spread mandated by our labor rules. This leaves us some potentially difficult decisions. The simplest solution here could be to use some trippers. But the Labor Agreement only allows a maximum of %, in effect meaning two of our eight. That would still leave us with two blocks not yet assigned to runs. One of the lessons we are learning here is that runcutting small solutions is often the most dif- ficult of the runcutting tasks that can be undertaken. Going back to the math, there are simply fewer options for cutting and hooking pieces in a manner that will allow a feasible solution. For some small runcutting problems there may in fact only be one feasible solution possible. So what are our options at this point? Let's note them and discuss the pros and cons: Create four trippers, one for each of the peak blocks . As noted above this is the sim- plest procedure but fails to meet the % maximum rule. There may be an opportunity to negotiate with the union and allow some variation to the labor rules, or to let one or two runs through that don't meet the rules. This is not the preferable option but at times may be the most pragmatic, particularly if faced with tight timeframes. Revisit the schedules and blocks, potentially reducing the spread of some of the blocks to allow the split runs to fall within the -hour spread limit. Or at least adjust one of them this way and leave the other two peak blocks as shorter part-time pieces. This can be an expedient approach for runcut purposes but may affect desired service level outcomes. Revisit the entire process and recut the pieces to allow "better" runs to be cut, and a legal solution to be created. This approach, assuming it allows us to find a solution, is often preferred as it allows us to meet our runcut objectives without affecting the schedule or blocks. Revisit the blocking to allow blocks that will cut into runs in an efficient manner. This is of course the approach that will take the most time and could affect the schedule itself. Another key lesson we are learning is that the schedules and blocks cannot be written in iso- lation from the runcut process. To do so is to ignore the fact that scheduling is an integrated and holistic process that requires all aspects to be considered. The scheduler who undertook the blocking would have given an indication of the costing impacts of the proposed schedule, 5-31

OCR for page 207
Chapter 5. Runcutting Level . Basic Runcutting based on the blocks developed. Right now those costs are probably completely inaccurate as the runcuts are not able to be provided efficiently with the initially developed blocks. run guide For the sake of expediency we will assume that in this case we are allowed one split run to go A summary of runs that describes over the -hour spread maximum. In practice this is not an option for the scheduler but for start/finish locations, work hours, now we will make this assumption as it allows us to see how the process unfolds. This will allow and cost element breakdowns. The us to create one split run and two trippers from the four peak blocks. Run Guide is the principal docu- ment that describes all of the runs How does one choose which is the split and which will be the part-time blocks? In this case we available for bid. choose the combination of blocks that minimizes the spread of our split run--Blocks and . Blocks and then become two trippers. Well done--we now have a completed runcut. Our run guide now has a completed runcut and is provided below. Option - Completed Runcut Start End Run Block Report Sign Off Plat Report Travel Total Spread Type Spread Guarantee Overtime # # Time Time Hours Hours Hours Hours Penalty Time Place Type Time Place Type 1 Str 2 5:46 6:01 Garage Pull 12:09 A Street 12:34 6:08 0:30 0:10 6:48 6:48 1:12 0:00 0:00 2 Str 2 11:44 12:09 A Street 18:38 Garage Pull 18:53 6:29 0:30 0:10 7:09 7:09 0:51 0:00 0:00 3 Str 3 5:35 5:50 Garage Pull 12:39 A Street 13:04 6:49 0:30 0:10 7:29 7:29 0:31 0:00 0:00 4 Str 3 12:14 12:39 A Street 18:19 Garage Pull 18:34 5:40 0:30 0:10 6:20 6:20 1:40 0:00 0:00 5 Str 5 6:05 6:20 Garage Pull 13:09 A Street 13:34 6:49 0:30 0:10 7:29 7:29 0:31 0:00 0:00 6 Str 5 12:44 13:09 A Street 19:53 Garage Pull 20:08 6:44 0:30 0:10 7:24 7:24 0:36 0:00 0:00 1 5:31 5:46 Garage Pull 9:19 Garage Pull 9:24 7 Spl 8:21 0:50 0:00 9:11 14:18 0:00 0:35 2:09 6 14:46 15:01 Garage Pull 19:49 Garage Pull 20:04 8 Pt 4 5:50 6:05 Garage Pull 9:38 Garage Pull 9:53 3:33 0:30 0:00 4:03 4:03 0:00 0:00 0:00 9 Pt 7 15:05 15:20 Garage Pull 19:19 Garage Pull 19:34 3:59 0:30 0:00 4:29 4:29 0:00 0:00 0:00 54:32 4:50 1:00 60:21 5:21 0:35 2:09 Note that in the above spreadsheet there are as many formulas as possible, holding true to our previously stated aim of avoiding typing numbers that can be calculated in every possible circumstance, no matter how complex the formula. 5-32

OCR for page 207
Level . Basic Runcutting Chapter 5. Runcutting Now we need to carefully check that we have all trips covered, all runs are legal (except for the agreed Split Run), that our numbers add up, and that the totals are calculated correctly. You should have a checklist and templates to assist with this process. One solution is to go back and review our graphical depiction of the blocks. For each block, "color in" or trace over the pieces and runs. This way we can be clear that the runcut covers the blocks entirely. The table below summarizes the total cost of the runcut. Runcut Summary Solution 1 Straight Runs 6 67% Split Runs 1 11% Total FT Runs 7 78% Part Time Runs 2 22% Total Operators Required 9 Hours Breakdown Total % Avg Platform 54.5 90% 6.1 Report 4.8 8% 0.5 Travel 1.0 2% 0.1 Total Work Hours 60.4 6.7 Penalties % Avg Spread 2.2 3% 0.2 Overtime 0.6 1% 0.1 Guarantee 5.4 8% 0.6 Total Paid Hours 68.5 Pay/Plat Ratio 1.2553 5-33

OCR for page 207
Chapter 5. Runcutting Level . Basic Runcutting As the table shows, we have seven full time and two trippers. The total weekday paid hours are . , and we have a ratio of driver pay hours to total platform hours (pay/plat ratio) of . . So is this a "good" runcut? To answer this question go back and review against the original stat- ed objectives. Clearly there is some inefficiency in the runs--over % of the cost is incurred in guarantee, spread, and overtime penalties. But these are a function of the constraints of the runcut (the rules and the blocks we started with) more than a reflection of the competency of the runcut itself. We would also need to look at historical trends for the route and the agency to understand if the solution fits in within normal bounds. Reliefs No job is over until the details are done. There are still a few things that need to be completed before the runcut can be considered final (assuming the agency decides that this is the runcut to implement). A key task relates to reliefs. The rules stated that all travels to relief locations must use a car. Generally this will mean that an agency has a car or fleet of cars that operators use to get to and from relief locations. We need to be sure we have enough cars for the travels dictated by the runcut. To be sure this is the case each travel required by the runcut should be listed--in ef- fect a list of "trips" that need to be scheduled into a "block" to be operated by the car (or cars). Car Trips Driven By Leave Arrive Driven By Leave Arrive Run # Garage Relief Run # Relief Garage 1 11:59 12:09 2 12:09 12:19 3 12:29 12:39 4 12:39 12:49 5 12:59 13:09 6 13:09 13:19 In this case the one car will be able to undertake each of the relief trips required. However, there may be cases when multiple reliefs occur at the same time, or at different locations, and more than one car could be required. In such cases a separate scheduling task is required to match cars to reliefs (in the above manner) to obtain car vehicle requirements. 5-34

OCR for page 207
Level . Basic Runcutting Chapter 5. Runcutting paddle An output of the scheduling process Information and Outputs that provides the operator with information regarding his or her Now it is a matter of generating the appropriate outputs and distributing them accordingly. workday--what time the work day Most transit agencies have their own specific set of reports or outputs for distribution. There- starts/ends, how to get to/from relief fore we will not provide samples here, but just list those that are typical. locations, the trips to be operated, and times at all timepoints. If an Run Guide. The run guide displays a summary of each run--where it starts, finishes operator drives on more than one etc., along with information about the length of runs. It may also include a breakdown route in the day, the paddle will of hours and penalties. This report is used across the organization as a means of quickly have all trips shown sequentially, as reviewing runs. well as travel paths between routes Paddle. The paddle comes in many shapes, sizes, and formats. The aim of the paddle is if needed. The paddle may also to provide the operator with information regarding his or her workday--what time the include a list of route turns, route work day starts/ends, how to get to/from relief locations, and the trips to be operated maps, farebox, headsign, and radio (complete with times at all timepoints). codes, and key intersections and Headway Sheet. We include this in runcutting as it may be helpful to operations staff at stops that must be announced. this stage if the run number for each trip is included. run number The unique number assigned to Automating the Runcut each work assignment on a specific The above process used spreadsheets to develop a runcut. However the majority of transit day.* At some systems, the run agencies now use computerized scheduling software packages to generate runcuts automati- number is unique only when used cally. in combination with a designator for the garage or the route or route Why bother with manual techniques then? As stated previously it is necessary to understand group number. the manual techniques in order to understand, and be able to review, the solutions provided by *Throughout this chapter, we have automated scheduling systems. Many schedulers make the mistake of setting up basic param- simplified the examples by refer- eters and then "pressing the button" to produce a completed automatic runcut. ring to Run 1, Run 2, etc. Obviously, these numbers are not unique. In Our experience is that it is extremely difficult to model all of the subtleties in a format (i.e., a practice, agencies typically use logical or mathematical statement) that will allow a computerized system to fully automate. multi-digit run numbers that can include the route or route group The more likely outcome is that solutions will be produced that achieve most of the aims of a number, a garage identifier, a code runcut but require some manual intervention or tweaking to finalize. for weekday/Saturday/Sunday, and a code for time of day. Each of these Great care must be taken to define all rules, constraints, and preferences to meet the runcut run numbers is unique on any given day. objectives. Take our above example as a good case. Here we were able to agree to allow one "bad" run with a spread above the maximum allowed. However if we used an automated sys- tem, with a hard limit of hours set, it would not have been capable of generating the solu- tion we created manually. This is an example of where the skill and experience of the scheduler are required to assist the automated tool to generate an outcome. In this example we may 5-35

OCR for page 207
Chapter 5. Runcutting Level . Basic Runcutting have needed to manually create and then force or fix the run before resubmitting the runcut. Only then could the computerized system have generated the same solution. The computerized system will significantly enhance the ability of the scheduler to produce runcuts if the scheduler is aware of how to make the most of the system as a tool, and not as a solution itself. Computerized systems also significantly enhance the final parts of the pro- cess--error checking and report production--to provide a level of automation and accuracy not otherwise possible. Other Factors Before ending the basic section, it is worthwhile to cover a few qualitative issues that can sometimes get lost in the quest for efficiency. The first issue is operator fatigue. Even in the absence of rules limiting spread time, it is a good idea to ensure that an operator's day is not too long. The scheduler's biggest concern regard- ing fatigue is to make sure that rostered operators get at least eight hours off between work shifts, whether or not this is spelled out in the labor agreement. Another issue is how "optimal" runcuts are defined. Cost, number of drivers, and number of buses are all extremely important factors, but poor run structure (and poor rostering) can cre- ate morale problems, leading to higher turnover rates and increased recruitment and training costs. These factors could eventually outweigh any cost efficiency benefits. A good scheduler will balance working conditions and efficiency factors in the runcutting process. 5-36

OCR for page 207
Level . Basic Runcutting Chapter 5. Runcutting LEVEL End of Basic Runcutting. 1 The Intermediate Section of Runcutting continues on the next page. To jump to Rostering, go to page - . 5-37