Abstract
Measuring cost effects of stockout events is a crucial aspect in Supply Chain Management. Motivated by the limited number of related approaches and an obvious need for applicable methods, our article focuses on this research topic. We present an interactive approach for obtaining cost function estimations from a decision maker. As a result of previous experiments, a knowledge-based Genetic Programming approach for rough function approximations is presented first. After that, a novel algorithm for reaching better convergence of the roughly approximated cost functions is described. Constituting the central research topic of this article, different interactive strategies for questioning a decision maker are then implemented and computationally tested. A decision support system integrating both aspects is illustrated and numerical results are presented. It is shown that the quality of computational results first depends on the quality of decision maker's answers. Second, looking at different questioning strategies in detail, cost function convergence seems to be reached faster, if decision maker statements are added to the database step- and not group-wise.