In this paper, we consider the integrated planning of resources in a service maintenance logistics system in which spare parts supply and service engineers deployment are considered simultaneously. The objective is to determine close-tooptimal stock levels as well as the number of service engineers that minimize the total average costs under a maximum total average waiting time constraint. When a failure occurs, a spare part and a service engineer are requested for the repair call. In case of a stock-out at spare parts inventory, the repair call will be satisfied entirely via an emergency channel with a fast replenishment time but at a high cost. However, if the requested spare part is in stock, the backlogging policy is followed for engineers. We model the problem as a queueing network. An exact method and two approximations for the evaluation of a given policy are presented. We exploit evaluation methods in a greedy heuristic procedure to optimize this integrated planning. In a numerical study, we show that for problems with more than five types of spare parts it is preferable to use approximate evaluations as they become significantly faster than exact evaluation. Moreover, approximation errors decrease as problems get larger. Furthermore, we test how the greedy optimization heuristic performs compared to other discrete search algorithms in terms of total costs and computation times. Finally, in a rather large case study, we show that we may incur up to 27% cost savings when using the integrated planning as compared to a separated optimization.

No abstract

We study a problem faced by a service provider, who is responsible for the repair of a group of assets subject to random failures. In case of a failure, both an engineer and a spare part of the right kind need to be available to carry out the repair. A limited number of engineers are employed while also stocks of the various spare parts are limited. In case any resource (engineer or spare part) is not immediately available, the service provider may follow a full backlogging policy. Alternatively, in case of spare parts stock out, he has the option to revert to an emergency supplier with ample capacity of resources. We present an original model to analyze the problem dynamics between this service provider and the emergency supplier. Especially, we determine the optimal emergency shipment cost and the optimal multi-resource level of the service provider. To this end, we propose a computationally efficient algorithm to find the Stackelberg equilibrium. Furthermore, we design revenue and cost-sharing cooperative contracts between these players which always result in coordination. Finally, we examine the risk of uncertainties in these contracts and find the optimal contract parameters by considering the utility functions of the players.

No abstract

No abstract