A Hybrid Model based on Genetic Algorithm and Space-Filling Curve applied to Optimization of Vehicle Routes

Abstract

This work is the result of a real problem in the Sanitation Company of Espirito Santo (Companhia Espirito Santense de Saneamento), which owns a Geographic Information System, but lacks a mechanism to build routes to server customers that open in average 2148 services requests per day. Therefore, we propose a Hybrid Optimization Algorithm that combines Genetic Algorithm and Space-Filling Curves to solve the Vehicle Route Problem. We establish the validity of the hybrid algorithm by performing tests in two different benchmarks datasets. Our proposal reached an average result of 12.7 percent and 4.1 percent better than the previous solutions in the first and second datasets respectively. Also, we compare our solution and five other variations of Ant Colony Optimization Algorithm. The results show that our proposal is superior in some simulations and, when it was not superior, presented the second-best results for almost all instances.

Keywords

• vehicle routing
• genetic algorithms
• fractals
• hybrid intelligent systems
• computer applications