Development of a hybrid genetic algorithm for multi-objective problem for a vehicle routing problem

A large number of logistic problems for Supply Chain Management are discussed in recent production environment. Most logistic problems are resolved as problems involving a single objective function such as total traveled distance, cost, and so on. However, different multiple objective functions are required in some real problems and their functions have trade-off relationship. Then, several objective functions are required to evaluate simultaneously and to search for Pareto solutions. This study discusses a vehicle routing problem in which different objective functions are evaluated simultaneously. And, hybrid type of genetic algorithm is proposed to resolve multi-objective vehicle routing problem. When different methods are adaptive to search for different single objective functions in the multi-objective problem, Pareto-front sought by either method tends to incline toward the axis of the objective function to which the method is more adaptive. In order to search for approximate optimal Pareto-front, the different methods which are adoptive for different single objective functions are incorporated into multi-objective genetic algorithm as the proposed method. The proposed method is examined on a bench-mark problem to evaluate its performance. The result shows that the proposed method obtains approximate Pareto solutions distributed uniformly in a coordinate system of objective functions in a practical computational time.