In the context of freight transportation in urban areas, the aim of this study is to show that making coalitions between parties (or carriers) involved in close areas, sharing transportation requests, vehicle capacities and warehouses, can increase vehicle utilization rates and reduce cost and emissions caused. Arising from the difficulties in finding solutions for the Travelling Salesman Problem (TSP) in feasible time, clustering methods have been lately used in order to reduce the computational time and provide more accurate results. In this project, the TSP is studied by creating different coalitions among the parties (or enterprises) involved, in order to reduce the total cost of them. Three different algorithms are explained, evaluated and simulated in this report. Two of them (DBSCAN and Spectral Clustering), are based on clustering methodologies. Once the coalitions have been formed and TSP has been resolved, one new important issue arises: which is the fairest allocation of the profit among the carriers so as to maintain the stability of the alliance? In this study, we also propose several cost allocation methodologies, describe its properties, compare and evaluate them. In detail, we study the following: Shapley Value, the Nucleolus, the Equal Charge Method, the Alternative Cost Avoided Method, the Cost Gap Method and the Equal Profit Method. All the study is based with several numerical experiments on randomly generated instances.

Outgoing