Multiple Criteria Optimization of the Carpooling Problem

The paper presents a multiple criteria (MC) formulation of the carpooling optimization (CO) problem and a solution procedure that allows to solve it. The mathematical model of the MCCO problem includes two major sub-problems, such as planning of the routes and matching carpoolers (drivers and passengers). Different aspects, including: economic, social, technical and market-oriented are considered. The MCCO problem is solved with the application of an original computational procedure based on the multiple criteria genetic algorithm, called NSGA II and the solutions' analysis and review technique, called Light Beam Search (LBS) method. The former method allows to generate a set of Pareto optimal solutions, while the latter assists the carpoolers in finding the most desired compromise solution (common route and match between carpoolers). The results of computational experiments are presented. We find that solving the formulating carpooling problem in a heuristic manner is possible in reasonable time (C) 2019 The Authors. Published by Elsevier Ltd.