Comparative evaluation of some interactive approaches to multicriterion optimization

This paper describes the results of a laboratory experiment the purpose of which was to compare from a human decision-maker's point of view the performance of the interactive methods for multicriterion optimization recently developed by Geoffrion and Benayoun et al. (STEP-method), and a simple trial-and-error procedure for solving the multiple criteria problem. All the methods were tested on a time-sharing system using as subjects both business school students and managers recruited from industry. The experimental context chosen is a problem of planning production, inventory and labour for a fictitious company.