Interactive aspiration-level multiple objective linear programming methods: Theory and computational tests.

摘要

This research develops a method based on aspiration levels of a decision maker for interactive multiple objective linear programming. The method assumes an unknown pseudoconcave preference structure of a decision maker throughout the decision process and the decision maker's ability to select a preferred solution from p + 1 alternatives, where p is the number of objectives. In addition to presenting the supporting theory and algorithm, we perform a comparative study using a fictitious decision maker comparing our work with those of Dell & Karwan (1989) and Steuer & Choo (1983), which all ask similar questions of the decision maker and utilize Tchebycheff functions. The use of a Tchebycheff function is motivated by its ability to reach an optimum at a nonextreme point solution. The statistics collected in the comparative study provide insights into the nature of the algorithms and the behavior of the solution techniques with different categories of problem structure and different underlying utility functions.;Based on the understanding of the comparative study, a heuristic procedure (a two-phase interactive approach) for multiple objective linear programming problems is proposed by implementing the learning and convergence process. This research also presents the results of a laboratory test with a simulated decision maker.