Return to Decision Line Home Page
Return to DSI Home Page

INFORMATION TECHNOLOGY

Lance B. Eliot, Feature Editor, Eliot & Associates

Using Simulation As a Real Tool

by Lance Eliot, Feature Editor

Suppose that you wanted to set up a new manufacturing operation for a brand new product that your company intends to produce. Before you rush out to buy the production line equipment and begin to physically lay out the entire assembly line process, you would probably want to investigate various options. How many units can you produce per production line operation? Can you increase production by using different equipment or different processes?

All of these preceding questions are part of a "what if" type of analysis that you would want to perform and then discuss with other personnel involved in the manufacturing development issue. Of course, you could use paper and pencil to try and work through all of the various options, but imagine how long and arduous the effort might be via the use of such a simple tool.

Instead, if you are even just a bit computer savvy, you would be better advised to explore the various scenarios by making use of specialized software packages that are for simulation or modeling. Similar in notion to using a statistical package to help you compute statistics for a survey, a simulation package aids you when trying to model a process and can be invaluable to understanding the trade-off's between various options.

Earlier in my career, when I held academic positions, I wrote and used simulations for teaching and research purposes. For example, business school students learned about the operation of a business by playing a simulation of businesses competing against each other (wherein students made investment decisions and the computer then simulated industry activities and displayed the results of each companies' performance). From a research perspective, I attempted to increase the sophistication of such simulations by adding expert systems and other artificial intelligence technology to them.

Recently, I had an opportunity to revisit the current status of the simulation marketplace by attending the 1996 Winter Simulation Conference. In addition to observing the latest research related to simulations, I also had an opportunity to review the many simulation packages available from vendors and see how real companies are using simulation to their competitive advantage.

If you have a potential need to perform a simulation, I strongly recommend that you consider making use of a modern simulation package. The latest packages are much easier to use, require less programming expertise to set-up and run, and have dramatically come down in price to be reasonable tools for even small-scale simulation problems.

Real Problems Being Solved

I want to emphasize that the simulations I am referring to can be applied to a wide variety of real world problems. Though I began the column by discussing a manufacturing-related problem, keep in mind that simulations can be used for all different classifications of problems (and you don't have to be an engineer to use simulations and simulation packages).

The massive 1,527-page conference proceedings describes dozens of fascinating examples of simulations being undertaken by companies, such as:

  • Simulations for business process reengineering. The user of the simulation package describes the internal company processes and then explores ways to reduce the time required to perform an activity or reduce costs associated with an activity.
  • Simulations for the heath care industry. The user of the simulation package describes a hospital work flow to determine the proper number of doctors and nurses per shift to handle the anticipated volume of patients.
  • Simulations for networking. The user of the simulation package describes a telecommunications network and explores the best ways to configure the equipment to adequately cope with data and voice traffic patterns.
  • Simulations for a call center. The user of the simulation package describes a call center operation and tries to determine the delays that callers might experience due to the staffing loads of the call center.

In each of the examples about the use of simulations, note that the developer of the simulation identified a specific problem that needed to be addressed, collected initial data about the nature of the operation to be simulated, acquired and learned how to use the simulation package, created the simulation in the simulation package, and then ran simulations to explore solutions to the specific problem being studied.

The sequence of steps required to properly make use of a simulation package must be carefully carried out. If you create a simulation based upon faulty data, the simulation results will be suspect. Likewise, if you don't understand how to correctly create the simulation, then the effort consumed to make use of simulation packages will be misspent.

Important Tips on Simulations

When examining the simulation packages that might be used to create your simulation, here are a few tips on what to watch out for.

Animation. In order to readily explain the simulation to other participants involved in your study, you may want to incorporate animation into your simulation. Thus, if you are analyzing a call center, you might show graphic icons of phones on the computer display and indicate the phones being answered as calls come into the call center. You could use colors, such as green for call completed and red for call abandoned, and otherwise make the simulation visually attractive to help other personnel understand just what the simulation is trying to do.

Some of the simulation packages allow animation, while others make it nearly impossible to incorporate animation into your simulation. Look at the features of the simulation package and carefully review the animation capabilities to make sure that they match your needs.

Provided Templates and Examples. Many simulations are really variations of other simulations. Rather than inventing a simulation from scratch, you should always try to determine if an existing model can be adapted for your current modeling needs. Is there already a telecommunications network model that you could revise to match your current telecomm simulation problem? And so on.

Some of the simulation packages provide templates and example simulations that you can then adapt for your own purposes. Also, some vendors encourage their simulation package users to share their models or contribute to a library of models. Thus, besides looking at the features of the simulation package itself, find out if any templates or examples are provided that might be readily revised to meet your needs.

Programming versus Diagramming. In the good old days (which weren't necessarily that good!), we had to write a simulation by making use of arcane coding commands. Today, most of the simulation packages allow you to diagrammatically create your simulation and then fill in forms that ask for information such as volume counts, probabilities, etc.

If your simulation requirements are straight forward, the diagrammatic approach should be suitable. But, if your simulation requirements are more complex, you may indeed need to augment your simulation with "programming" -- thus, determine whether the package you are reviewing will allow both diagrammatic and code programming.

Trying out the Packages

Many of the vendors provide demo disks containing their simulation package or offer access to their simulation package via their Web site (or, some provide a 30-day trial period whereby you buy the package and can return it if you are dissatisfied). I tried quite a number of the packages to understand their capabilities and be able to see just what the latest packages can do. Here are a few of my favorites.

Simul8 (Visual Thinking Inc.). This PC-based simulation package is one of the least expensive ones (under $500) and yet has a number of surprisingly impressive features. The animation capability lends itself to quickly animating a model, and the tie-in to Visual Basic made it easy for me to expand my models. For those circumstances where you want to quickly put together a simulation, and need to keep your costs down doing so, Simul8 is a great choice.

Extend (Imagine That! Inc.). Another PC-based simulation package, but moving up closer into the $1,000 dollar club, Extend takes a somewhat scientific approach to model building and offers various libraries of simulation "blocks" that you bind and build together to create your simulation. Extend also encourages programming via the language ModL. Using Extend will require some devoted attention, and assumes that you are quite serious about being into simulation.

Micro Saint (Micro Analysis & Design Simulation Software Inc.). Moving further up into the pricing range, near to $10,000 dollar packages, Micro Saint had some interesting examples of using simulations for an army tank control panel layout, branch banking customer flow, and others.

Conclusion

Deciding to use a simulation is a much easier choice today than even 5 years ago. Prices for the simulation packages are now at several target levels (under $500, near $1,000, near $10,000, above $10,000). And, with advances in graphic displays, animation, diagrammatic programming, and other ease-of- development features, anyone seriously interested in conducting a simulation can find a way to do so.

So, if you are a business analyst or even a manager, and have a problem that might lend itself to simulation, you might want to explore your options about which options you can explore. Take a look at the simulation marketplace and you might like what you see.

Remember that your input is welcomed. If you have projects addressing the information technology area, and you would like to share this with the readers of "Information Technology," please contact me.