On the Value of Information for Tactical Decision Support in Stockout Management
Sebastian Langton, Martin Josef Geiger · 2012
This article presents an approach for decision support in tactical stock out management, primarily with respect to inventory management problems. A framework for eliciting information on stock out consequences from experts is presented. In a subsequent evaluation process, our solution approach for detecting cost functions in soft expert information is introduced: By applying a Genetic Programming (GP) algorithm, sufficiently precise cost functions can be determined for analyzed stock out items. Comparing the information content of different test instances, the quality of solutions seems to be satisfying even with little information available. Furthermore, it is shown that a tendency of positive correlation between the amount of information in test instances and the quality of the solutions exists. Finally, an obvious gap of solutions quality can be observed at a certain degree of instance-specific information content.