Exact Formulation and Analysis for the Bi-Objective Insular Traveling Salesman Problem

Abstract

This paper aims at studying the Bi-Objective Insular Traveling Salesman Problem (BO-InTSP), which searches for a set of efficient, single visit sequences to collect (or distribute) freight from a set of islands. In this problem, the selection of ports (nodes) to be visited at each island, along with the associated port visit sequence, are optimized simultaneously, while the maritime transportation costs and the ground transportation costs inside the islands are minimized with a bi-objective perspective. This approach is employed since these costs are of a conflictive nature. A previous Approximated Formulation of the BO-InTSP relies on aggregating the actual demand locations within each island in a certain number of centroids for computing the ground transportation costs. Conversely, this paper proposes and develops a novel Exact Formulation for the problem based on the actual demand locations, instead of aggregating the demand inside the islands. Additionally, a systematic evaluation approach is developed to compare the two alternative formulations with different levels of demand aggregation inside the islands, considering the bi-objective nature of the problem. The results reveal that the novel Exact Formulation significantly outperforms the previous aggregated approach in terms of the solutions quality and computational resources.

Keywords

• insular traveling salesman problem
• ground transportation costs
• freight collection or distribution
• isolated regions
• bi-objective optimization
• multi-objective analysis