Comparing Genetic Algorithms and Simulated Annealing for Solving the Pickup and Delivery Problem with Time Windows

Solving the Vehicle Routing Problem (VRP) and its related variants is significant for optimizing logistic planning. One important variant of the VRP is the Pickup and Delivery Problem with Time Windows (PDPTW), where it is assumed that a client may request one of two service types, a pickup or a delivery. It is required to find a set of minimum cost routes for a fleet of vehicles, while observing a number of predefined constraints. In this research, we try to handle the difficult constraints using a new solution representation and simple neighborhood moves that will maintain the feasibility of solutions throughout the search. Our solution method is tried within two meta-heuristic approaches, a Genetic Algorithm and a Simulated Annealing. Based on the performance of the two algorithms on a number of benchmark problems, we draw conclusions about which among the two algorithms seems more appropriate for solving the problem.