A novel mathematical model and a large neighborhood search algorithm for container drayage operations with multi-resource constraints

Abstract

Multi-resource constraints in container drayage operations exist widely in real-life scenarios but have been seldom reported in literature. This study mainly addresses a container drayage problem with a limited number of empty containers at depots, with the goal of minimizing both the number of trucks in operation and the total working time of trucks. The problem is firstly formulated as a bi-objective mathematical model based on a so-called determined-activities-on-vertex graph. Three schemes are proposed to handle the developed mathematical model as: (a) a nonlinear constraint is linearized; (b) a parameter that is a large enough constant in the model is analyzed and tuned deeply; and (c) the bi-objective model is converted into a single-objective model. Furthermore, a large neighborhood search algorithm is designed to solve the problem. The two solution methods are validated and evaluated based on randomly generated instances as well as instances from literature. Numerical experimental results indicate that the methods can provide optimal or near-optimal solutions for medium- and large-scale instances in a short running time.