The repair kit problem with imperfect advance demand information

To promote customer loyalty and generate revenue in after-sales service, many companies provide field repair services to their customers. Service technicians who perform these repair jobs typically carry a set of spare parts called a repair kit in their company van. The repair kit problem aims to determine which parts to include in this kit and in what quantity. Currently, many appliances are equipped with sensors that monitor their different functionalities. If an appliance breaks down because one of these functionalities is disturbed, then the respective sensor triggers a failure code that describes the appliance’s condition. From the service technician’s perspective, this failure code serves as potentially imperfect advance demand information for spare parts. In this paper, we present an extension of the repair kit problem that uses this information for the replenishment decision. We formulate this repair kit problem with advance demand information as a Markov decision process and propose two heuristic solution procedures. Our first heuristic is far-sighted and optimizes the inventory of all parts individually, while our second heuristic is a myopic greedy algorithm that considers all parts at once. We conduct an extensive numerical study to evaluate the performance of both heuristics and to identify which heuristic performs best under which circumstances. Comparing both heuristics to a state-of-the-art algorithm that disregards any available advance demand information, we find that utilizing this information yields substantial cost savings.