Abstract
Order based combinatorial optimization problems usually require decomposition to work effectively. Such a methodology is implemented in Dynamic Programming, Tabu Search and other meta-heuristics in the area of VRP and has proved to be successful. However, it has not been tested in Genetic Algorithms. Here, we develop a new meta-heuristic, called the Localized Genetic Algorithm (LGA). We apply this LGA to small scale capacitated vehicle routing problems, both with and without route constraint. Our results on popular datasets have demonstrated competitive performance.