Decision-Based Design Using Time-Varying Preferences Represented by Stochastic Processes

Soliciting and expressing the preferences of a decision maker in engineering design is critical. In general, the preferences vary through time, complicating the design of engineering systems. In this article, we propose that if parameterized utility functions are used to model the preferences, the time-varying characteristics of the parameters can provide valuable information on the likely decisions the decision maker can make at a future time. To model the time-dependent uncertainty in preferences, we use parameterized utility functions with the parameters characterized by stochastic processes and demonstrate how the design process is affected by stationarity properties of the random parameters. We work in the normative utility theoretic domain and show a property of the multiplicative utility function that allows us to use the common Black-Scholes-Merton options pricing model from finance, to account for variability in preferences with time. Finally, we discuss how to modify the design process so that optimal products are ready when there is a future need for them. The applicability of our approach is demonstrated using a cell phone example.