A HYBRID GENETIC ALGORITHM FOR OPTIMIZING URBAN DISTRIBUTION OF AUTO-PARTS BY A VERTEX ROUTING PROBLEM

Abstract

The present work designed and implemented a hybrid algorithm by combining the genetic algorithm meta-heuristics (GA) with the nearest neighbor algorithm (NN). We combined these algorithms to solve the Capacitated Vehicle Routing Problem with Time Windows and a single depot (CVRPTW). The proposed implementation optimizes the distribution for an auto-part trading company within the urban perimeter of Quito city – Ecuador, by designing a script coded in C# language. Besides, to evaluate the quality of the solutions generated by the proposed hybrid algorithm, different inst ances of the problem were built by taking small samples from the whole customer's information. To compare the performance of our algorithm, we used a VRPTW model encoded in GAMS. In addition, we applied the problem from the case-study company in real instances. As a result, the generated sequences of the routes travelled by the trucks reach an improvement of close to 20%. We calculated that percentage using the Euclidean metric.