Utilização de algoritmos genéticos híbridos e Web 2.0 no desenvolvimento de sistemas de roteirização de veículos

Abstract

The use of a computer system for vehicle routing is essential for the competitiveness of companies working with distribution of resources and products. Despite high cost, such a system ofiers a guaranteed return on investment made. Using a Metaheuristics techniques combined with the resources currently ofiered by web, free development tools can be developed routing systems with features, performance and eficiency very close to commercial systems. In this work, was first performed a study of the theoretical foundations of Genetic Algorithms. Using the Java programming language, is implemented a set of routines common to the AGs, in the form of a library of development. With this routines, was developed an algorithm routes of vehicles with support for multiple routes and various types of calculations of costs between locations. To improve the performance of the algorithm implemented, we used the technique of local search 2opt, and are thus resulting in a Hybrid Genetic Algorithm. This algorithm was used as a core for the development of a functional prototype of a web vehicles routing system capable of handling the geographical coordinates of settlements in land surface and create routes on a map of vehicles considering the vehicular trafic routes . The system also generates a detailed list of routes and close to describing direction to be taken on each site to facilitate the work of a driver of the vehicle.