A sorting based efficient heuristic for pooled repair shop designs. (May 2020)

• Record Type: Journal Article
• Authors: Turan, Hasan Hüseyin; Sleptchenko, Andrei; Pokharel, Shaligram; ElMekkawy, Tarek Y
• Is Part Of: Computers & operations research. Volume 117(2020)
• Publication Date: 2020-05
• Subjects: Maintenance -- Repair Shop -- Pooling -- Queuing -- Heuristic; Operations research -- Periodicals; Electronic digital computers -- Periodicals; 004.05
• DOI: 10.1016/j.cor.2020.104887
• Languages: English
• ISSNs: 0305-0548

Abstract

Highlights: A joint problem of repair shop design, inventory and capacity optimization is solved. A sorting based efficient heuristic is developed. The proposed sorting heuristic achieved less than 1% average optimality gap. Around 4% cost reduction is achieved compared to the best benchmark method. The pooled repair shop designs based on service rates are cost-efficient. Abstract: In this paper, we address the assignment problem of skills to servers (e.g., repairmen) in a multi-server repair shop of a spare parts supply system. This type of assignment problems tends to be hard in general, due to the lack of analytical queuing models with skill-based item-server assignments. In this paper, we propose a joint skill-server assignment and inventory optimization heuristic based on "pooled" repair shop designs. The heuristic decomposes the repair shop problem into sub-systems based on some attributes of repairable items. Each subsystem is responsible for its group of repairable items with full cross-training of the subsystem servers. The pooled designs reduce the complexity of the problem and enable the use of queue-theoretical approximations to optimize the inventory and repair shop capacity. The conducted numerical experiments show that the pooled skill-server assignments optimized by the proposed heuristic can reduce the total costs by 4% when compared to the skill-server assignments obtained by Genetic Algorithm and Simulated Annealing based methods. Furthermore, in terms of cost and computation speed, the proposed heuristic shows better results than a Simulation-Optimization based skill-server assignment heuristic, which considers all possible assignments.