Solving Dynamic Rail-Car Fleet Sizing Problem by a Hybrid Metaheuristic Algorithm

The aim of this paper is to present a new solution method to optimize rail-car fleet sizing problem. The model is a dynamic and multi-periodic and car demands and travel times are assumed to be deterministic. There are important interactions between decisions on sizing a rail car fleet and utilizing that fleet. Consequently, using empty rail-cars is considered in the proposed model that leads to a high decrease of fleets number and costs. The model also provides rail network information such as unmet demands, number of stationed rail-car in each station and each time period, and number of loaded and empty traveling rail-car at any given time. In this paper, the proposed solution procedure is combining Genetic Algorithm (GA) and Simulated Annealing (SA) algorithms. GA is used for determining the number of fleets in each station at the beginning of the planning horizon, and SA is used for assigning rail-cars to demands in the length of the planning horizon. Computational results demonstrate the applicability of the proposed approach. For evaluating the proposed solution method, a comparison between results of the hybrid metaheuristic algorithm and optimum results of CPLEX software is done. The efficiency and effectiveness of the proposed hybrid algorithm is revealed.