بهینه سازی یک شبکه بازسازی تصادفی با یک گزینه مبادله

An international manufacturer of industrial equipment offers its customers a remanufacturing service consisting of a refurbishment of the most critical part in order to rejuvenate the equipment. Offering remanufacturing services is in line with a servitization strategy. We develop a strategic decision support tool to optimize the required remanufacturing network. Investment decisions have to be made, not only concerning the number and locations of remanufacturing facilities, but also concerning the appropriate capacity and inventory levels to guarantee specific service levels. These network decisions are influenced by the way remanufacturing services are offered. We consider two service delivery strategies, either a quick exchange of the used part by an available remanufactured one or re-installing the original part after it has been remanufactured. Given the high level of uncertainty, we build a stochastic, profit maximizing model to simultaneously determine the optimal network design and the optimal service delivery strategy for a multi-product, multi-level network for repairable service parts. The rapid modeling formulation with a non-linear objective function subject to non-linear constraints is solved by the differential evolution algorithm. We conduct the analysis for fast and slow moving part types. The model can be easily extended to more general settings, while the case-study provides valuable insights for practitioners.

The large installed base of equipment provides original equipment manufacturers with an opportunity to develop a profitable remanufacturing business. The launch of more extensive warranties and overhaul services relies on remanufacturing activities and fits the current servitization trend in the industrial equipment industry [4], [19] and [5]. Consequently, an increasing number of companies like Bosch and HP are intensifying their remanufacturing activities [10]. Our case-study company is an international manufacturer in the compressed air and generator industry with a renewed focus on remanufacturing. Due to confidentiality reasons the company is referred to as AirGen and financial specifications are omitted. In order to set up a remanufacturing network, AirGen's management has to decide upon the number and locations of facilities. In addition, appropriate capacity and inventory levels have to be set in order to fulfill the service level agreements (SLA). Furthermore, contractual arrangements made with customers (e.g. regarding the ownership of parts) and the selected service delivery method have an impact on the optimal remanufacturing network design. The goal of our research is to build a model that supports this complex decision making process at the strategic management level.

In this paper we have taken a profit maximization perspective to optimize the remanufacturing network design and the delivery strategy for a global manufacturer of industrial equipment. The choice of the service delivery strategy, refurbishment either with or without an exchange option, has implications on costs and revenues of the network. Apart from this decision and the profit perspective in the developed model, our contribution lies in the simultaneous solution of three related network design problems, i.e. the facility location problem, the capacity and the inventory problem. For this company's multi-echelon repairable network structure, we determine the optimal number and location of remanufacturing facilities. Taking into account the stochastic nature of both demand and processing times, a queueing model is integrated into the model to set proper capacity and inventory levels, while a specific customer service level is guaranteed. Due to the non-linearity of the rapid modeling formulation, we solve the model by applying a differential evolution search procedure. The case study results emphasize the importance of volume as a driver for decentralization. As volume goes up, the number of remanufacturing facilities increases accordingly. Sales price is the main driver for the choice of the service delivery strategy. Next, the choice of the service delivery strategy influences the capacity levels in the optimal network structure. In addition, the impact of increased service levels will differ depending on which service delivery strategy is chosen. Therefore, we emphasize the need to simultaneously analyze the design of networks and the mix of service contracts. Slow moving parts should be remanufactured in a central facility, whereas fast moving parts tend to be more decentralized. Note that the recommendations resulting from this research have been fully implemented by the case study company. The authors acknowledge that the insights derived from the case study need further support by more case study research and sensitivity analysis. Nevertheless, our research shows that a holistic approach in network optimization is necessary. Several extensions are possible e.g. taking into account transportation batching, investigating limited field technician capacity and allowing for non-Poisson distributed demand. However, by assuming general distributions for both transportation and remanufacturing times and by allowing multiple part classes, multiple resources and multiple network echelons in the model formulation, the decision support tool can be easily applied to any general case.