Integrated optimization of finished product logistics in iron and steel industry using a multi-objective variable neighborhood search

Abstract

This paper is concerned with the integrated optimization of finished product logistics that consists of two sub-problems: consolidation planning and transportation scheduling. In this problem, three kinds of decisions should be made simultaneously, namely the selection of candidate finished products in each warehouse (i.e., consolidation planning) and the vehicle allocation and transportation sequence of selected finished products in each warehouse (i.e., transportation scheduling). Since the loading times of products to vehicles are sequence-dependent, a tradeoff should be made between two conflicting objectives: maximization of loadage of ships and maximization of efficiency of the whole logistics. In practical industry the schedulers usually handle the two sub-problems separately, which often cause much lower efficiency of the whole logistics and higher transportation cost. Therefore, in this paper the two sub-problems are integrated and formulated as a multi-objective mixed-integer programming model, and then a two-layer multi-objective variable neighborhood search (TLMOVNS) algorithm is developed to solve it. Computational results on simulated instances show that the proposed TLMOVNS is very efficient for this problem and much superior to the current scheduling method generally used in practice.