Optimization by integrating engineering and business models

Abstract

To optimize is to find the "'best" solution given certain conditions and constraints. What is meant by best and how to find it has received scant attention. To the engineer, "best" may be fastest, strongest, most reliable and so on. The engineer will have models to determine whether one solution is better than another based on objectives such as these. "Best" in business is quite different. It means to maximize profit or minimize loss. The economist or accountant also has models. Note the obvious disconnect between the objectives of engineering and business models. This disconnect has hindered us from finding a practical best to improve profit on the plant floor. Simple questions like "what is the best tension to run" have no useful answers from a strictly engineering or business viewpoint.

This paper begins by defining best for several familiar examples. However, it quickly concludes that the only "best" that makes sense in an industrial environment is that which will minimize total costs. To find this best we must integrate engineering and business models. This technique developed here is very powerful, flexible and adaptable approach. The technique can be applied explicitly using calculus or similar numerical techniques when cost functions are well known. Even more flexible is an implicit approach which can be used when very little is known about costs. Five web handling examples are used to illustrate this problem solving technique. These examples include a variety of objectives such as optimum rejection levels, core waste, web tension, and layon roller nip and water flow rate. These examples show how it is easy to combine apples and oranges, such as waste and delay, when one converts to a common denominator of cost.