The added value of advance demand information for the planning of a repair kit

In this paper the repair kit problem is studied, where technicians have to visit several customers to repair broken appliances (such as copiers or heating systems) and they can only take a limited set of parts with them (called the repair kit). In this problem, it has to be decided which spare parts to include in the repair kit. We consider a version of this problem in which partial advance demand information is available. That means we divide the set of parts into two subsets, where the condition of parts in one subset is monitored by sensors. In case an appliance fails and a repair job is requested by a customer the service provider is able to access this sensor data before a technician visits the customer. For this setting, we derive an expression for the job fill rate, which is used as a constraint in the optimization model, where holding and replenishment costs are minimized. We use a greedy heuristic to determine near-optimal repair kits. In a numerical study, we find that integrating advance demand information yields substantial cost savings. In order to find out for which parts having advance demand information is most valuable, we examine the effect of parts' demand probabilities and their prices. We find that monitoring parts that are expensive and likely to fail leads to the largest cost savings. In particular, the price of the monitored parts and the achievable cost savings are strongly correlated.