August 1996 € Volume 23 € Number 4

Second Look at Simulation Software

Non-traditional uses can yield unexpected benefits

By Jerry Banks and Van Norman

The traditional uses of simulation have been discussed by many authors [Banks and Norman, 1995; Law and Kelton, 1991; Pegden, Shannon and Sadowski, 1995; and Schriber, 1991]. However, there are other non-traditional uses of simulation software that produce unexpected benefits. This article presents an expanded set of simulation uses that are commonly considered, plus some non-traditional uses and their benefits. The listing was developed by examining many past simulation projects.

Traditional uses
The traditional uses of simulation go beyond providing a look into the future. These include the following:

Better understanding. Managers often want to know why certain phenomena occur in a real system. With simulation, you determine the answer to the "why" questions by reconstructing the scene and taking a microscopic examination of the system to determine the cause or causes of the phenomena. This is difficult to accomplish with a real system because you frequently can't see or control it in its entirety.

Also, by compressing or expanding the time scale, simulation allows you to speed up or slow down phenomena so that you can thoroughly investigate them. You can examine an entire shift in a matter of minutes if you desire, or you can spend two hours examining all the events that occurred during one minute of simulated activity. Time compression and expansion helps the decision maker to understand "why."

Animation features offered by many simulation packages take this a step further by allowing the user to see the facility as if it were running. Depending on the software used, you may be able to view the simulated operations from various angles and levels of magnification, even 3-D.

Visualizing system dynamics adds to understanding and aids in the detection of design flaws that appear credible when seen just on paper or in a 2-D CAD drawing.

Explore possibilities. Simulation lets you test every aspect of a proposed change or addition without committing resources to their acquisition. This is critical, because once the hard decisions have been made, the bricks have been laid or the material handling systems have been installed, changes and corrections can be extremely expensive.

Simulation can be used to specify requirements for a system design. For example, many firms use simulation to define complex control algorithms and then implement the rules in the actual system. Additionally, the specifications for a particular type of resource in a complex system to achieve a desired goal may be unknown. By simulating different capabilities for the resource, the requirements can be established.

Additionally, a great advantage of using simulation is that once you have developed a valid model, you can explore new policies, operating procedures or methods without the expense and disruption of experimenting with the real system. Modifications are incorporated in the model, and you observe the effects of those changes on the computer rather than the real system. Simulation, in many firms, is integral to the continuous improvement process.

The cost of a typical simulation study is usually less than 1 percent of the total amount being expended for the implementation of a design or redesign. Since the cost of a change or modification to a system after installation is so great, simulation is a wise investment.

Diagnose problems. The modern enterprise is very complex; so complex that it's impossible to consider all the interactions taking place at a given moment. Simulation allows you to better understand the interactions among the variables that make up such systems. Diagnosing problems and gaining insight into the importance of these variables increases your understanding of their effects on the performance of the overall system.

Long throughput times, missed production targets, late jobs and other production problems keep many managers awake at night. It's easy to forget that congestion is an effect rather than a cause. However, by using simulation to perform bottleneck analysis, you can discover the cause of the delays in work-in-process, information, materials or other processes. Once the cause is determined, an appropriate solution can be tested via simulation and implemented.

Build consensus. Some people operate with the philosophy that talking loudly in meetings, distributing fancy reports and presenting computerized output convinces others that a proposed design is valid. In many cases these designs are based more on an individual's preconceived notions than on factual analysis. Simulation provides objective understanding about how a system actually operates -- a much firmer basis for design than relying on subjective opinions.

In these days of downsizing and cut-backs, many people are fearful. Using simulation to make objective decisions can help to alleviate these fears. Since decisions are made objectively, employees are less likely to be drawing any sort of inference when you approve or disapprove of designs.

Train the team. Simulation models can provide the basis for excellent training. Used in this manner, the team makes decisions at set points during the progression of the model. The team, and individual members of the team, can benefit by their mistakes, and learn to operate better. This is much less expensive and less disruptive than on-the-job learning.

When Not To Use Simulation
To be fair, we should point out that there are times when simulation is not an appropriate tool. In fact, if a closed-form mathematical model is available, simulation should not be used since it takes more time and is probably more expensive. Also, because simulation models are inherently descriptive they offer little in the way of prescriptive advice concerning how to manage or modify operations. Consequently, if a different type of model (linear programming, for example) can be used, it is more appropriate to do so.

For instance, if we are trying to plan production levels for the next year, a detailed, minute-by-minute, simulation of the entire factory is probably inappropriate. A linear program using aggregate production capacities would not only provide "what if" information, but would also suggest "what's best." The problem with most prescriptive models is that they often do not accurately represent the complex interactions and dynamics of the real-world system.

Non-traditional Uses
There are many uses of simulation that are non-traditional. Oftentimes, these non-traditional uses may be more important than the traditional ones.

Modeling the system contributes to understanding. In a complex system, no one person knows all the details. The modeling effort benefits everyone's understanding of the unfamiliar components, the relationship of the components to the entire system, and the impact of an individual sub-system on the entire system.

In a manufacturing environment, one individual might be acquainted with metal cutting operations, whereas another understands the material handling devices, and yet another is an expert only in the information flow system. If all of these people are participating in the simulation activity, either as modelers or through inclusion in the review sessions given by the modelers, they become more familiar with other components of the system. Many opportunities for improvement arise when all involved can see the big picture.

An example from consulting pertains to a warehousing and distribution center that was not achieving the desired throughput. In this case our client was another consulting firm serving as the system integrator. The simulation model indicated that the resources were not being used anywhere near their limits. The problem seemed to be in the algorithm that controlled stock picking.

The system integrator continued to provide us with revised picking algorithms, but none were significant improvements. In desperation, they asked us, the simulation analysts, to devise an algorithm since we had developed a very thorough understanding of the system. The algorithm developed by our analyst was a significant improvement over that developed by the system's integrators, although it still did not achieve the desired throughput.

Data collection leads to understanding. There is often so much noise in raw data that it is difficult, sometimes impossible, to "see the forest for the trees." Looking at histograms and other graphical output can provide a much clearer picture of what is happening. For example, an understanding that the repair time for a resource follows a distribution that is skewed to the right might help understand why long waiting lines sometimes arise. Further, this understanding might lead to some alternatives for reducing these long waiting lines, even before the model is complete, or the simulation analysis has taken place.

During the construction of a simulation model on a recent consulting engagement, we were trying to match our model's predicted output with the output of an actual heat treating process. Unfortunately, the output of the model greatly exceed the actual output. We had accurate data on the time between batches from an automatic recording device and had assurances that the center ran two machines after a major shutdown. Because of the discrepancy, we spent several months collecting actual output data.

Eventually, we realized that although the operation was running two furnaces as we had been told, the output was low the first few days after a shutdown due to restart problems. This fact had never been entered into the model (or into capacity planning, for that matter). When added to the model, the predicted output was in complete agreement with the actual and we learned something significant about the system.

Visualize the idea. Consider the case in which there are disagreements between managers concerning the desirability of a course of action. One manager says to go straight, another says go to the left, and the third says go to the right. Once a manager commits to a direction, it is a show of weakness to change to the direction championed by another manager. An impasse can be reached, and no action is taken, perhaps to the detriment of the firm. Simulation can avoid this impasse, or even break through it if it has been reached. After verification and validation a simulation model can become reality. This reality can then be used to resolve disputes as well as become a specification for the manner in which the system will be operated.

Not only does simulation help with the visualization by managers, but with hourly workers and supervisors that must be convinced that an idea has merit. In fact, it is quite possible that a system will fail unless all of those involved buy into it. New methods may be perceived as risky and threatening to hourly personnel. Some of this fear is warranted. The workers wonder if they will be able to make 100 parts per day under the new system. They wonder what will happen if they can only make only 80 parts per day. These workers may become resistant to change.

Process debugging. Oftentimes, when validating a model of a rather complex manufacturing facility, the simulation appears to run significantly better than the actual facility. Usually this is considered a "bug" that needs to be corrected. However, if the "bug" can be replicated in the actual facility, this may be an opportunity for improvement. We refer to this as "process debugging."

This has happened to us on several occasions. Once, this discovery led to a significant reduction in throughput time. The model assumed that a single cart was used to move work from a department having many parallel machines. The resulting cycle time was 30 percent less than the actual time, and we looked for the "error." When we found that each of the parallel machines had its own cart and made the appropriate change in the model, the cycle times matched. Later, we realized that the actual department could function with a single cart. Thus, the "error" in the simulation study became a recommendation that was unexpected.

Process debugging can lead to big opportunities for improvement in the actual system. A simulation model makes it easy to "debug" the process by finding out which assumptions are being violated and what the impact will be by improving them. The simulation model also pinpoints the changes that will make the most impact. [For more on process debugging, see Hopp and Spearman, Chapter 9. ]

References
1. Banks, J., and V. Norman (1995), "Justifying Simulation in Today's Manufacturing Environment," IIE Solutions, November.

2. Hopp, W.J., and M.L. Spearman (1996), "Factory Physics: Foundations of Manufacturing Management," Richard D. Irwin, Burr Ridge, IL.

3. Law, A.M., and W.D. Kelton (1991), "Simulation Modeling and Analysis," 2nd ed., McGraw-Hill, New York.

4. Pegden, C.D., R.E. Shannon, and R.P. Sadowski (1995), "Introduction to Simulation Using SIMAN," McGraw-Hill, New York.

5. Schriber, T.J. (1991), "An Introduction to Simulation Using GPSS/H," John Wiley, New York.

Jerry Banks is a professor in the School of Industrial and Systems Engineering at the Georgia Institute of Technology and associate director for research at the Center for International Standards and Quality at Tech. Van Norman is president of AutoSimulations, Inc. in Bountiful, Utah.

For more information about this article, put the number 2 in the appropriate space on the Reader Service Form

E-mail to the Editorial Department of OR/MS Today: orms@lionhrtpub.com

Lionheart Publishing, Inc.
2555 Cumberland Parkway, Suite 299, Atlanta, GA 30339 USA
Phone: 770-431-0867 | Fax: 770-432-6969
E-mail: lpi@lionhrtpub.com