Combined Logistic Regression and Zionts-Wallenius Methodology for Multiple Criteria Linear Programming

摘要

The use of logistic regression to infer the values of weights in a linear utility function representing the decision maker's preferences is incorporated within an interactive method of the Zionts-Wallenius type. This provides a method with much of the simplicity and power of the Zionts-Wallenius method, and with additional robustness to deviations of preference statesments from consistency with a strict linear utility function. The method also provides for making explicit use of indifference as well as preference statements. Application of the method to a small problem is discussed.