ACHIEVING TOTAL SIMULATION
We're at the point now, I think, where most engineers work with computer-aided design/manufacturing (CAD/CAM) system on a daily basis. Originally meant to speed design and reduce reliance on paper, CAD can now do 3-D visualization of a fully rendered design. The dynamic characteristics of a product or process are also handily dealt with on the most modern platforms. What CAD is to products, simulation is to processes.
The technical envelop for simulation is being pushed by factors found outside the realm of manufacturing systems. Once breakthroughs are made, though, it doesn't take long before the latest developments are applied to the manufacture of goods.
The military recently demonstrated before congressional committee a virtual, real-time, nationwide, battlefield simulation. Many types of simulations contributed to the results and small segments of "reality" were included as well. The program's purpose is to train personnel and to test tactical response and new weapons concepts. Reality segments provide the "unexpected response necessary to validate process."
Computer games sold to consumers generate so much cash that the industry is willing to invest megabucks in product development. Consequently, computer games have become tremendously sophisticated in a relatively short period of time. Simulation of evolution, the management of a city, or the behavior of an ant hill trains player to interact with representations of reality, i.e., reality as process.
How can this notion of games as a model of reality be applied to manufacturing? Let's design a "SimMan" game that allows the players to simulate manufacturing processes such that interactions are possible faster-than-real-time. Bringing the game to bear on the design, testing and manufacturability of products will enable lust-to-dust analysis of the product life cycle.
Before we begin, what are some of the problems with simulations to date? One big problem is validation. Not one manufacturing scheduler has so far gained wide acceptance in industry. This is simply because random events and sporadic occurrences spoil the best-laid plans of both mice and men. The simulation process needs to be certifiable. Speed of interaction should never be an issue. The simulator should be faster than what is modeled. Sometimes I think one of the big advantages the Japanese enjoy is that they don't apply MRP II widely.
What are the benefits of playing such a game? "What if" questions can be answered quickly. With a city simulator, the mayor can be mayor for over 200 years within a span of several month. (Who say that time-travel isn't possible?). For manufacturing, costs in product and life cycles can be ascertained. Automotive companies can weigh the expense of quick model changeovers against the marketplace advantages to be gained thereby. Agile manufacturing and small-lot concepts can be reality-checked.
Today, we continue modeling with hard and fast assets. In other words, too often we build it and then we think it through. First the factories are build and the products are made, then come the surprises. By then, changes are difficult if not impossible. Further down the road, responses to changes in market forces become expensive and time consuming.
Therefor, the simulation/game must prove valid over a wide range of possibilities. It has to be played in simulated time, at least an order of magnitude faster than the process itself. It should be extensible into marketing and aftermarket situations. This interactive model should examine hard and soft costs changed to the manufacturing process.
Using the military or games simulation techniques, we aim at lust-to-dust scaling. Validity is attained by inserting pockets of reality throughout the model. The pockets occupy about 1 percent of the model and might include segments of the manufacturing and supplier structure.
Connecting real running machines to the simulator increases the validity of the modeling process. Via distribution across a wide geographic and management landscape, scheduling gets a reality check. Scallability allows players to test small segment of the manufacturing process or expand into the design and marketing at a later date.
Of course, the final step will be from simulation to reality. It's been recently suggested that the line between and control is becoming blurred. In this near future, simulation is control. Not only will what if questions be answered, but simulation will also serve to clarify how we can "make it so".
As appeared in Manufacturing Systems Magazine March 1993 Page 60
References - Table of Contents
|R. Morley Incorporated
614 Nashua Street, Suite 56
Milford, NH 03055-4992 USA
Tel: 603-878-4365 FAX: 603-878-4385