The Future of OR

Lee W. Schruben is a professor and former chair of Industrial Engineering and Operations Research at UC Berkeley, and one of the world’s leading authorities on simulation theory and practice. He has consulted widely in the high-tech and biotech industries, as well as in banking and in auto making. He spoke to BNET about the challenges ahead for OR and for those in the OR field.

BNET: What are the big unsolved problems for OR in practice today?
Schruben: We have to realize that what we’re actually doing is forecasting. We are trying to model what will happen in the future, and that’s the biggest practical challenge — how to get away from models with static assumptions and develop predictive models that can respond in real time to changes in the world.

BNET: What do you mean by static assumptions?
Schruben: As it is now, we collect data and build a model based on assumptions we think are reasonable at the time. But our results only tell us what would have happened in the past, when those assumptions were valid. Most models assume that input data are independent and identically distributed (IID), but that’s almost never true. Assuming IID data is assuming that events in the world don’t depend on each other, and that the probability of them happening doesn’t change over time. But things are changing constantly in business.

read more The Future of Operations Research | BNET
 
Top