Optimization Models for Inland Haulage Transportation of Containers

In container intermodal transportation, a significant portion of the total cost arises from the inland transportation of containers. There are many parties (shipping lines, haulage companies, customers) sharing this operation as well as many restrictions that increase the complexity of this problem and make it NP-hard. Shipping lines and haulage companies tend to apply efficient optimization techniques to manage this process in away to reduce the overall cost and to ensure that customers are satisfied. In this thesis, we focus on container inland transportation from the perspective of delivering 20ft and/or 40ft containers on a heterogenous fleet of trucks, between a single port and a list of customer locations and inland depots. We investigate three types of inland transportation problems: Homogenous Container Sizes, Heterogenous Container Sizes and Stripe and Discharge of Containers. Each of the above problems has its own complexity but all have been classified as NP-hard problems. For this reason we will study these problems separately and the main contributions are describing, modelling, solving and analysing of the: • Homogenous ContainerSizes: an efficient assignment Mixed Integer Linear Programming (MILP) model is formulated which solves large scale instances in a reasonable solution time and can be implemented on variants of the container drayage problem. • Heterogenous Container Sizes: a Mixed Integer Linear Programming (MILP) model for combining 20ft and 40ft, Stripe orders is designed, which solves more efficiently than its previous analog. For realistic instances, a decomposition and aggregation heuristic is designed and tested to be cost saving. • Strip and Discharge of Containers: a Genetic Algorithm (GA) approach is designed and tested for solving large scale problems within a quick computational time and the result shows that combining the Strip and Discharge types with the usage of inland empty depots is cost and fleet saving.