Multi-size container transportation by truck: modeling and optimization

This paper investigates a multi-size container truck transportation (MSCTT) problem in which a truck can carry one 40 ft or two 20 ft containers. This problem considers both fixed and flexible drayage orders. A state-transition-based method is presented to formulate the MSCTT problem. The problem is modeled as a sequence-dependent multiple-traveling salesman problem with social constraints in which the distances between cities depend on the sequence of cities visited before. Three tree search procedures and an improved reactive tabu search (IRTS) algorithm are designed to solve the MSCTT problem. They are validated and evaluated based on randomly generated instances extensively. The IRTS algorithm can solve small instances to optimality within less than 1 min. Furthermore, it outperforms the existing reactive tabu search algorithm and can solve realistic-sized instances efficiently. Several interesting findings about the MSCTT problem are presented. Finally, the IRTS algorithm is also applied to a real-world data set.