Emerging Information Technologies for Facilities Owners: Research and Practical Applications: Symposium Proceedings (2001)

An industry-sponsored program for implementing promising technologies in the construction industry and a study that estimated the economic gains achievable through their use were outlined at the conference by Dr. Richard H.F.Jackson, Managing Director of the FIATECH (Fully Integrated and Automated Technology) Consortium, and Dr. Robert E.Chapman, an economist with the Office of Applied Economics at the Building and Fire Research Laboratory at the National Institute of Standards and Technology (NIST). New technologies for managing the design of fast-track, capital projects were discussed by Dr. Ray Levitt, Director of Stanford University’s Center for Integrated Facility Engineering. Dr. Sarah Slaughter, President of MOCA Systems, Inc., described a model of construction activities that reintroduces the link between design and construction. Mr. Rick Hendricks of the General Services Administration reviewed how an information technology application was used to plan for the relocation of the U.S. Patent and Trademark Office from 34 leased locations to a new 5-building campus.

Summary of a Presentation by Richard H.F.Jackson Director, FIATECH Consortium

The construction industry is being challenged to build and maintain facilities more rapidly, at less cost, that are sustainable, safer, more integrated, more performance driven, and more flexible. The industry needs both basic technology research with long-term horizons and applied research that helps deploy existing and emerging technologies more effectively and more imaginatively. The FIATECH (Fully Integrated and Automated Technology) Consortium was formed in late 1999 to accelerate the deployment of advanced technologies that will improve profitability in the construction industry.

FIATECH is an industry-led, collaborative, not-for-profit research and development consortium launched under the auspices of the Construction Industry Institute and in cooperation with the National Institute of Standards and Technology (NIST). It currently has more than 40 members, including some of the largest chemical and pharmaceutical firms; many of the leading architect-engineering, computer-aided design and manufacturing software firms; the U.S. Army Corps of Engineers; and NIST’s Building and Fire Research Laboratory.

The FIATECH vision is to bring owners, operators, contractors, and suppliers together from all across the

industry to achieve this vision of full integration and automation. The problems involved are too complex for any one organization. Competition has taken on global dimensions. The basis of the competition has broadened from price to quality, to time to market, to innovation, to design in dexterity of supply chains. It is only by working together that we can get to the seamless integration of information flow and from all participants throughout the entire project life cycle, from concept to design, to construction, to operation, to maintenance, and even to decommissioning and dismantling.

The consortium is modelled after the highly successful National Center for Manufacturing Sciences and SEMATECH, a group formed to reinvigorate the semiconductor industry. FIATECH will operate by forming, managing, and deploying complex, multi-partner research, development, and deployment projects. It will have a full-time staff that will develop leveraged funding opportunities and manage the technical projects, including:

• seamless integration of information flow among all participants throughout entire project life cycle;

• application of the latest available proven technologies;

• computer-aided drawing and computer-aided engineering;

• advanced communications;

• field sensing and tracking;

• field automation;

• construction automation;

• breakthrough improvements in quality, schedule, and cost, resulting in improved return on investment;

• pre-construction; and

• modularization.

The construction industry has not been able to benefit fully from these technologies, largely because it is a highly fragmented industry with a short-term project orientation and low research and development (R&D) budgets. It spends less than one percent of its revenues on R&D.

There are few broad industry standards. There is no common industry voice to deal with development. Even if there were, some of the costs in these areas are too high for any one company to bear. There is no common vision for full integration and automation. And there is yet no common roadmap to guide us.

It is our intention in the FIATECH Consortium to address these roadblocks head-on and by doing so, to help achieve reductions of 30 to 40 percent in cost and schedule time. These improvements will reduce design changes and rework through concurrent engineering and better control of project scheduling and cost, improve supply chain management, detect differences between designer intent and construction, develop highly accurate as-built information for operation, maintenance, and renovation, and ensure that the right data are available when and where they are needed.

FIATECH has a Board of Directors to whom I report. Strategic Focus Areas, or SFAs, will be the essential operating units of the consortium and will guide most of the consortium’s activities. SFAs will function as focus areas or interest groups within FIATECH.

Each SFA will be led by an elected board and charged with developing high-level goals and objectives, along with a strategy for accomplishing them. Membership dues will be used to provide administrative support of most SFA activities, but SFAs will usually seek sponsors for projects. One important sponsor is likely to be the federal

government, which spends $500 million a year in construction-related research. Project participants will cover the cost of the research and deployment, which will allow for differential allocation of intellectual property that provides the profit motive for company participation. We expect some projects and products to be made generally available, but the participants will decide what will be disseminated freely and what will be proprietary, licensed, or sold.

The five members of the former Owner/Operator Forum (Air Products and Chemicals, BASF Corporation, Dow Chemical Company, DuPont Company, and Merck & Company) recently joined FIATECH en masse. They will be managing the consortium’s first strategic focus area: defining business drivers and requirements for software related to the life cycle of capital equipment.

We are also starting a development cycle for several other projects. We will be working with the Construction Industry Institute to identify a project related to electronic commerce. We are talking with the Army Corps of Engineers and NIST about an effort that will help support the Corps of Engineers’ desire to integrate their projects.

Interoperability is a very important issue. FIATECH is considering development of cost-effective technologies for collecting, compiling, and maintaining field data for actual representations of buildings. These would include advanced sensing and scanning tools to collect the data, wireless technology for moving the data where they are needed, and visualization software for providing meaningful representations of the data and analysis software to ensure you get what you want.

Another idea is that FIATECH should operate a test-bed for fully integrated and automated project processes. This would be a place where researchers and engineers could work together, either collocated or not, to test, evaluate, try out, and demonstrate in a low-risk environment, new ideas, software tools, and best practices for information technology integration. NIST already has this capability and is ready to work with us to put it to use in areas of mutual interest.

In such a test-bed, high-quality interactive simulation tools might be used for developing three-dimensional models of plants and facilities, for checking out operating procedures for new facilities, and for gauging the impact of introducing new technology and conducting performance analysis. A design team, owners, and contractors might come together and immerse themselves in a proposed design to help determine whether it is really what they are looking for. Owners and operators could also use the facility to walk around in both designs and as-builts and determine the differences.

Ways of exploiting information technology in the construction industry are often lumped together under the rubric of electronic commerce. It is important to distinguish between e-commerce, which may be defined as buying and selling on the Internet, from e-business, where you have business-to-business data sharing, requests for proposals, invoices, bidding, some project management, and so forth.

There is also a notion of e-construction, which is actually connecting the front office with information and data and what is going on all the way down to the construction site using radio-frequency identification tagging, bar coding, or global positioning system monitoring.

Much progress has been made in the top two layers by application services providers, but little has been done in the bottom, infrastructural layer. And even in the upper layers progress has been slow because of the lack of industry standards. It turns out that few people and few organizations invest much in standards, although we talk about their importance. Yet it is at that bottom layer that the foundation for progress is laid.

Summary of a Presentation by Robert Chapman Economist, Building and Fire Research Laboratory, National Institute of Standards and Technology

Fully integrated and automated project processes (FIAPP) technologies are characterized by one-time data entry, interoperability with design, construction, and operation processes, and user-friendly input and output techniques. If implemented, they would result in significant reductions in both the delivery time of constructed

facilities and life-cycle costs of those facilities. And cost and cycle time reductions are vital if the construction industry is to remain competitive.

To promote the timely delivery of FIAPP products and services to the construction industry, NIST’s Building and Fire Research Laboratory (BFRL) formed the Construction Integration and Automation Program. This program is a interdisciplinary research effort within BFRL—in collaboration with the Construction Industry Institute, the private sector, other federal agencies, and other laboratories with NIST—focused on the development of key enabling technologies, standard communications protocols, and advanced measurement technologies.

The role of standards is crucial in the implementation of these technologies. NIST is uniquely positioned to take on that role, because it is specifically enabled to develop these infratechnologies—technologies that allow private sector individuals and companies to build their own proprietary technologies on top of it. This contrasts with the role of the FIATECH Consortium, which is more a vehicle for delivering results to the industry.

The stakeholders, in the effort to develop FIAPP, are all the key players in the construction industry: building owners and managers; building materials, equipment, and software providers and innovators; building contractors; and providers of support services. When FIAPP products and services are available commercially, construction industry stakeholders will benefit from reductions in first costs, reductions in delivery time, reductions in maintenance and repair costs, improvements in construction safety, and higher contractor profits. Offsetting these gains are the costs of these new technologies and increased research and development costs.

First costs of construction projects will be reduced because the technologies will result in less rework: better work scheduling, better work packages, being able to do simulations that will allow tasks and crew activities to be compressed and thus reduce delivery (cycle) time. Having electronic as-built information will reduce operation, maintenance, and repair costs. And having the ability to conduct simulations will have a favorable impact on construction safety.

Also we see an opportunity for productivity improvements. Manufacturing productivity has improved over the past 20 years; construction productivity has not. Improved construction productivity and better schedule controls by contractors translate into higher profit margins for contractors.

To quantify these benefits and costs, I authored a microstudy published by NIST in June 2000. The study, Benefits and Costs of Research: A Case Study of Construction Systems Integration and Automation Technologies in Industrial Facilities, is available on the World Wide Web at <www.fire.nist.gov/bfrlpubs/build00/PDF/b00025.pdf>.

As the title of the study indicates, it focuses on one segment of the construction industry—industrial facilities. This segment is the smallest of the four segments composing the construction industry—industrial, commercial, public works, and residential—accounting for less than 10 percent of the total value of construction put in place. However, the industrial facilities segment of the construction industry exhibits a number of characteristics that will in all likelihood lead to both earlier and more rapid adoption of FIAPP products and services. First, it is less fragmented than the other segments of the construction industry and tends to be dominated by larger construction companies. Once large construction companies adopt an innovative technology, they are able to leverage their subcontractor tiers, leading to increased levels of adoption throughout the supply chain. Second, the companies who own industrial facilities tend to be large in size, often exhibiting extensive use of information technologies in

their core business functions. Thus, they are better able to evaluate the merits of FIAPP products and services and encourage their use by contractors throughout the facility life cycle. Finally, many large industrial companies have actively pursued strategic alliances to promote increased cooperation and reduced confrontation between owners and contractors.

The study finds that the use of NIST’s FIAPP products and services is estimated to produce a $2 billion savings nationwide to industrial facility owners, managers, and contractors in the period from 1993 to 2015. The estimated savings are in 1997 dollars calculated at a 7 percent real discount rate. Ranges of 1 to 4 percent reduction in first costs and reduction in cycle time of 12 to 18 percent are projected.

An additional analysis was conducted using information on projected ranges of savings and costs. This analysis concluded that savings to the industrial facilities segment of the construction industry of $3 to $4 billion—expressed in 1997 dollars—were feasible.

Summary of a Presentation by Raymond Levitt, Ph.D. Center for Integrated Facility Engineering, Stanford University

“The Virtual Design Team” (VDI) is a technology developed by the Center for Integrated Facility Engineering (CIFE) to better manage the design of fast-track capital projects. The center is a collaborative effort of Stanford University’s civil engineering and computer science departments and innovative companies interested in deploying information and construction technologies.

The commercial version of the technology, which was seeded by small amounts of money from the center and larger grants from the National Science Foundation, is offered commercially as “SimVision™” by Vité Corporation of Mountain View, California.

CIFE was encouraged to develop the technology by companies like Intel and Palm for whom time to market is crucial. For microprocessor vendors like Intel, opening a plant ahead of schedule can be worth $1 million an hour during the few months that new chips sell for premium prices. It is $12 million of profit per day and that market window lasts only about three to four months. The technology is now in common use in the semiconductor, computer, telecommunications, oil, and processing industries, and is beginning to be used in construction projects.

The major challenge of fast-track projects is their large overhead generated by parallel work processes. The critical path methods developed during the 1950s and 1960s model sequential activities, such as building a foundation and then building the structure, where you have sequential handoffs. They are inappropriate for fast-track projects where there is a great deal of parallel activity. With fast-track projects, people on the project team spend substantial amounts of time coordinating with one another, and more significantly, reworking efforts initially based on partial information, as more and new information becomes available.

Project managers tend to underestimate the resources involved in coordination and rework with fast-track projects. This is because the relationship between the degree of fast tracking and the resources is exponential, not linear. The more you fast track a project the greater the relative amount of coordination and rework required. So the question is, where are you on the exponential curve? Are you on the early part or the middle or are you at the point where the coordination and rework is actually larger than the direct work?

The VDT/Vité design tool allows organizations to predict the amount of coordination and rework as a function of the activities being overlapped, the skills of the people responsible for those activities, the structure of the organization, and so forth. The modelling tool works much like a critical path model except that it looks at work as a volume of information to be processed in parallel, instead of a series of sequential tasks.

We are focusing on knowledge, design, planning, procurement—what happens before construction. The participants can be viewed as information processors with some set of skills and experience that affect two things: how fast they can do the work and the rate of errors they incur doing it. The higher the skills and experience of the team or individual relative to the demands of the task, the faster they go and the fewer mistakes they make. The difference between a highly skilled architect or engineer and a novice can be dramatic.

Fast-track projects are information intensive. Typically they are used to produce high-performance, complex facilities with a high level of interdependency among its subsystems. The fast-track schedule triggers unplanned coordination and rework for the project team, which must process a large amount of information under tight time constraints.

What is happening on these fast-track projects is dependent on the skill and experience of the people doing the work and the structure of the organization—how decentralized or how centralized it is. Rework has to be done when changes are made that affect other people. All of this takes time and this is what delays fast-track projects.

So the information-processing capacity of a project team is really the critical limiting factor for a fast-track project. We obviously need the resources to do the direct work, but we typically make allowances for that based on experience. What we tend to underestimate are the coordination and the rework. The more aggressively we fast track, the bigger an issue this is. Complicating all this are the products that get more and more complex, and the subsystems that interact in more complicated ways.

Yet, we cannot over-design things anymore. The world is too globally competitive. So, we tend to have lean designs with very interdependent subsystems. Now we try to build those in parallel. Every change ricochets up and down the organization and it gets absolutely overwhelmed with information processing. The trick is to be able to anticipate it, plan for it, and take steps in advance to fix it.

How do you handle this need to manage organizational risk in fast-track projects? The traditional way was to learn by doing: trial and error. You set objectives, propose an organization and a work process. You try it, see what happens, and learn from it. Project organizations have trouble learning, because we disband them and then we form them again in different ways. The answer proposed by the designers of the VDT/Vité tool was to design a new project organization the way you design a structure or a foundation or an energy system, that is: set objectives, propose a solution, model or simulate in advance many different solutions to predict the outcomes, and then choose the one that has a good chance of being successful. Of course, it is critical that you have a theory and some sort of mathematical or computational modeling tool that allows you to make predictions about how things will perform.

Vité starts with business milestones and activities needed to achieve them. It adds to the critical path model two other kinds of relationships. The first is called information exchange. If we know that choosing the facade material and choosing the construction methods are highly related activities, we will show an information dependency. The individuals responsible for those two tasks would spend time talking in the real world, so we want to simulate that in the model.

The second relationship is called change or failure propagation. For example, you apply for an excavation permit and discover a problem that will require a change in the construction method. The model will capture statistically a prediction of the rework that might be necessary. This gives a more realistic estimate of how much information it actually takes to get the job done.

The next step is to model the project team and try to understand its information-processing capacity. We define the sets of skills needed for the project and classify the participants in terms of their skills. (McGraw-Hill’s Construction Information Group has developed a database that could be used in this application. Many organizations have internal databases they could adapt for this use.) People using Vité then model the reporting relationships on the project. The final item modelled is decision-making policies in the organization: How central is decision making? How formalized is it? How much experience is there on this team? Have they worked together

before? How strong or weak is the matrix? Are the specialists collocated or are they in functional silos in different parts of buildings and offices?

Essentially, the “engine under the hood” of VDT/SimVision™ is a discrete event simulation where we are passing pieces of information through the virtual actors; they decide what to do, they do it, they make errors statistically, and they communicate with other people. The result is a series of reports that have turned out to be very accurate over and over again. A simulation of Lockheed Martin’s attempt to develop a commercial space launch vehicle in a schedule compressed from the typical four to five years to one year predicted that the vehicle would launch four months late. It launched within two days of the time predicted by the model. One of the reports is an estimate of who in the project team will actually be backlogged (i.e., where the bottlenecks will be). This is important because the effects of backlogs cascade. As an example, take a project involving the design of a semiconductor facility where the architect is backlogged 26 days, or 5 weeks, in what should be a four- to five-month project.

The project is delayed because this person’s tasks are delayed. Second, other workers’ time is wasted because they have to wait to get answers to questions. Third, it affects project quality because when you are five weeks backlogged you keep your head down and do your own work and you give less priority to answering questions and going to meetings, and doing coordination activities that affect the quality of other people’s work. All of these result in predictions of risk of process quality, which have been shown to correlate well with product quality problems. By identifying the potential failures, you can find ways to resolve them before they actually occur.

VDT/Vité simulations tend to underline an important lesson. When projects are fast-tracked, the knee-jerk reaction is to add more people to the team. But instead of more people, you might want to put your A-team players on the project because skill reduces the error rate and speeds things up. Higher skill often has a bigger impact on quality than more people. More people will speed up the process but will not improve the quality of the product. So you can consider different kinds of alternatives in advance in a “flight-simulator” mode with this technology.

Often the VDT/Vité team models the efforts of 10 to 20 participants in a project with 30 to 40 activities during start-up meetings for a project team. We build the model. We identify the bottlenecks and then go through alternatives quickly. In the course of a day we can come up with substantial insight into project risk and then iterate alternatives.

In summary, the VDT/Vité methodology and software provide:

• a common language to discuss and understand fast-track project work processes;

• a methodology and software to (re)design and improve those processes;

• a medium to share the results intuitively; and

• a framework to facilitate continual improvement and dissemination of best project designs.

CIFE is now studying the use of the technology in service and maintenance activities, which represent a larger cost for facilities owners, but where work processes are less well-structured. It will require a more branching, dynamic model because you often do not know in advance what the activities are. Models developed for health care, which is people maintenance, appear to be useful in simulating facilities maintenance. CIFE is also working with a consortium sponsored by the National Science Foundation that includes the University of Southern California, the University of Illinois at Urbana-Champaign, and Carnegie Mellon University. The consortium will study the network organizational forms that link knowledge and people throughout organizations and people outside organizations as well, and explore how distributed knowledge may affect the performance of organizations in the twenty-first century.

Summary of a Presentation by E.Sarah Slaughter President, MOCA Systems, Inc.

One of the things I have found in working in design and construction is that even when innovation takes place, it is not captured and reused. So we are constantly reinventing the wheel. I have also found it is difficult to get people to commit to using an innovative technology or process because there is a high degree of risk and uncertainty. There is a growing gap between design and the realization of that design, particularly in regards to time and cost in the construction industry that is due, in large part, to the large number of complex systems and complex processes. We need a mechanism to tie them together.

When I was teaching at the Massachusetts Institute of Technology’s (MIT) Department of Civil and Environmental Engineering, my students and I conducted time-motion studies and interviews at more than 200 construction sites focusing on the physical components in buildings and the tasks that are done to transform and aggregate those components into finished systems. The National Science Foundation funded the research. We captured the detailed task elements in the language of the carpenters and other people working on the site (the subcontractors, general contractors); tied that in with the knowledge of the architects and structural and mechanical engineers; and finally with the measures that matter most to the owner. The information captured was used to develop simulation models (Models of Construction Activities) that can be used in the early design stages to evaluate design and construction alternatives, examine innovation opportunities, improve resource utilization, plan for constructability, and investigate cost/time trade-offs. Thus it reestablishes the link between a design and its realization in construction.

In 1999 I took a leave of absence from MIT to establish a firm, MOCA Systems, Inc., in Newton, Massachusetts, to commercialize the computer models. MOCA Systems now offers a Web-based service for owners and developers, and their project teams of general contractors, construction managers, architects, engineers, design-build professionals, and specialty contractors. The model allows users to simulate the actual construction of a building from the ground up as it actually would be done in the field. It provides people with a system that corresponds to their tacit knowledge and allows them to run controlled experiments. You can sit down during the early design stages, put in information about the project and get back how much it is going to cost, how long it is going to take, and what the exposure of workers is to dangerous conditions. It can also be used to experiment with alternative techniques that might be used, learn what the impacts might be, and thereby reduce risk and uncertainty.

One model was developed and calibrated during the renovation of the Baker dormitory at MIT in three and one-half months—something that was not possible using standard methods. In planning the project, the dormitory construction team came up with some clever approaches to prefabricate members, which reduced the duration of construction by one-third. These innovative approaches also reduced the cost of the project by 20 percent. The model estimated the actual construction time and cost within 1 percent accuracy.

In another case, approaches developed through the use of simulation identified ways to reduce project duration from 39 days to 27 days to totally renovate a 150-year-old school building. For example, we looked at two processes: walls and plumbing. On a typical construction project the carpenters start on the walls; then they stand around and wait for the plumbers to rough in all the horizontal plumbing for the fixtures. By simulating the process

we found that by doubling the number of plumbers, we were able to loosen that constraint and increase the utilization of those resources.

The MOCA Build System has two parts. The core technology, which is licensed from MIT, is a library of construction process information down at the level of every single bolt, every single beam, for every single worker. What the users have is their project data and in that project data they can put all sorts of information. What is the size of the building? How many stories? Is it steel or a concrete structure? Is it a glass curtain wall or pre-cast concrete panels? What is the interior fit out? Do you have hard walls or do you have movable partitions?

All of this information goes into the MOCA Build project analysis. Users can also put in recent bid experience. You can look at the time and cost. What is the exposure of workers to dangerous conditions? The model uses Occupational Safety and Health Administration categories of causes of injury for each task and scales it to how long a worker is performing that task. The index scales by the size and attributes of the project. So if people are carrying things upstairs and downstairs rather than using a lift you are able to capture and compare some of those elements.

The second part consists of individual models for structural steel, cast-in-place concrete and light wood framing, models for glass-metal curtain walls and pre-cast concrete panels, five service models, and two interior finish models. The models are linked so that the user can determine, for example, when the steel is set faster, whether to enclose the building earlier and begin roughing in the systems, or whether the systems should be roughed in from the bottom up or top down. In the early conceptual development stage of a project, the models provide a way to actually test alternative designs and construction techniques. Users find this a great way to test alternative approaches for realizing their program requirements and meeting their time and cost objectives.

During the bid process and once construction begins the model provides details on almost every physical component in the building, and that allows everyone to work from the same database. We can also go into incredible detail and be able to talk about who is actually working where and when. This is an invaluable tool for bidding and mobilization planning.

Optimizing the use of resources can be of great value where expensive resources, such as large cranes costing $5,000 to $10,000 per day, are required. Simulations are also useful where high performance is required or where site conditions are constrained. You may have difficulty getting deliveries. You may have difficulties in moving elements. You may not be able to put the crane anywhere, and in that case the high complexity of the processes themselves cannot be mapped with the human mind. The human mind is a wonderful element but it cannot follow secondary and tertiary impacts down through a complex flow. The model basically keeps track of many of those interdependencies and their impacts on all the secondary and tertiary activities.

Owners who want to improve the development of designs they will be replicating in multiple locations also favor MOCA simulations. Rather than starting from scratch every single time, they would like to have a template and be able to analyze and improve on that template incrementally over time, or to quickly change the template to adapt to site specific issues.

Release of object-based computer-aided design (CAD) programs and CAD programs based on industry standards such as aecXML will only make the MOCA Build models easier to use, because it will potentially enable direct connection between CAD programs and the MOCA Build models. During the next year MOCA Systems will be commercializing a number of research models. We are developing a Web-based interface for inputting data and looking at ways to import information from legacy systems of other vendors.

Summary of a Presentation by Rick Hendricks Project Manager, General Services Administration

One of the most difficult challenges faced by the General Services Administration (GSA) in overseeing the consolidation of the U.S. Patent and Trademark Office’s (PTO) operations into a new 2-million-square-foot headquarters in Alexandria, Virginia, was not the planning and managing the design and construction of the

buildings. Rather, it was planning the move itself, a move out of 34 leased locations in 18 buildings with myriad different lease expirations, renewal options, and other conditions.

A key objective of the joint move team overseeing the consolidation was to keep the payment of double rent, what the GSA calls “overlap,” to a minimum. To this end, they developed an extraction plan to guide them during the move, which will take place over a 40-week period starting in November 2003.

Traditionally, moves of this sort are planned by GSA with customer-agency (PTO) facilities management professionals and consultants. The problem with this approach is no matter how long you study, or how good a program you write, or how smart you are at getting to understand the client, you cannot learn everything about the client in a finite amount of time. So, the plan is inherently flawed. It becomes a no-win proposition. Clients want to be satisfied, but you can never satisfy a client unless you give them everything they want, and the nature of a project is that you cannot give them everything they want. There are too many trade-offs. The real solution then is optimization, not total satisfaction.

Hence, early on in planning the PTO move, the team decided to use a simulation model to test a variety of scenarios. In the process the team shifted its role from expert planner to facilitator and used their experience to keep the project on track and to avoid mistakes. To help visualize the move and engage all elements of the PTO the team decided to simulate the move by setting up the scenario-planning model as a game. Gensler, the space planning and interior architecture consultant on the team, was willing to try this approach and brought their expertise to the mix.

So far, the model has since been run six times and will be run several more times. The results have been better than expected. After using the model the people making decisions or recommendations understand that they are in an environment of trade-offs. They recognize that optimization rather than total satisfaction is the goal.

For instance, for those using the model it becomes quickly apparent that they cannot all have the top floor that faces the Washington Monument. Once they understand that not everyone can have the prime office space, they ask: What are we going to do? How are we going to resolve it? GSA no longer has to sell a solution, because the client understands how they arrived at the overall solution and they buy in to the process.

Team members intervene as little as possible in the scenario planning. In one scenario PTO managers proposed using vacated space still under lease as interim space while new space was prepared, because of the savings that could be achieved. If such a step was unilaterally proposed by GSA it would likely have been resisted. By using the model, everyone understands the rationale and the result.

The simulation also helps agency managers deal with the inevitable issue of the government paying double rent. Simulating different scenarios allows GSA to find out, for example, that $10 million spent on vacant space is much better than the $45 million that might have been spent if the move had been planned in the traditional manner. Without these realistic, quantifiable options, $10 million might look like failure, not optimization.

In running a scenario the objectives are to do so quickly, accurately, and realistically and to fully document the course of the game. The scenario-planning game is played in a large room with three tables. On one side of the room are tables covered with large schematics of the 108 floors in 18 buildings being vacated, and on the other side are the floors of the five large buildings to be occupied in the new complex. Poker chips with a label for each organizational unit in the PTO representing 10 people each are placed on the appropriate floor of the schematics for the existing building. A table on the third side of the room has schematics for expansion space and the requisite chips. (I originally considered using software that would have allowed the whole game to be played in a virtual fashion, but rejected the idea because the tactile game being used proved to be non-threatening, forgiving, and very engaging. It was a social event, like a family puzzle.)

The players and support people are briefed. Flip charts in the room spell out the overall strategy for the game, assumptions, lessons learned, and issues to be resolved. Professionals from PTO, GSA, Gensler, and other consultants are on hand to provide advice on technical issues. Team members are appointed to make sure that elevators and loading docks work within their capacities—a “loadmaster” in the old space, a “dockmaster” in the new space, and several “experts.”

The idea of the game is to devise the best way to move the chips from the schematic representing the existing space to the schematic representing the new space at the other end of the room, while taking into account when the newly constructed space will come online and when the old leases expire. Reality and complexity are introduced by also trying to satisfy adjacencies, operational continuity, and information technology needs. Everyone is cautioned to take into account the physical and time constraints. Using real managers subliminally introduces the more subtle and hard-to-define factors like corporate culture, turf issues, unspoken desires, and body language in a creative, manageable way.

Chip moves are devised, negotiated, and agreed on by the “players” (the PTO managers). The chips are manipulated by “croupiers” at the direction of the players during the course of the game. The chips are transferred to the new buildings by people designated as “drivers” who carry 35 chips (the amount of people who can be efficiently moved in a week) in each “moving van” (actually a foam-core carrier with a space for each chip.)

Digital cameras and laptops were used to record the moves in the game. Gensler enters the results into a computer system using a mix of software, primarily Aperture (a computer-aided design/facilities management system that will be used to manage PTO’s new space), Excel, and Powerpoint. The software allows the decisions made in the course of the simulations to be documented, displayed, and even replayed. Financial costs like rent, downtime, and move costs are computed and displayed.

PTO and GSA officials will soon be able to develop a comparative analysis of the completed scenarios to arrive at the best all-around solution. They will be using a decision-assistance package called Expert Choice. Participants will be brought together and asked to respond to a series of questions dealing with decisions made during the simulations and the perceived outcomes. Using Expert Choice, they will be able to spot trends, pitfalls, unanticipated consequences, and opportunities. It should assist the move team to rank the scenarios generated and to further optimize the moving process.