A Combined Approach to Solve the Pickup and Delivery Selection Problem

In this article we propose a model which decides upon the execution of transportation requests in order to maximize their contribution to the overall profit. Positive contributions of requests are obtained by composing appropriate routes for the vehicles involved. The composition of routes is hindered by limited capacity of vehicles and time windows associated to requests. A hybrid approach consisting of a parallel path construction heuristic which seeds a Genetic Algorithm is presented. We assess its capability for a set of suitable benchmark instances for pickup and delivery problems with time windows