Investigating Genetic Algorithms for Solving the Multiple Vehicle Pickup and Delivery Problem with Time Windows

One important problem in transportation and logistics systems is the Multiple Vehicle Pickup and Delivery Problem with Time Windows (MV-PDPTW). The problem deals with a number of pickup and delivery (P&D) requests that should be served by a fleet of vehicles. A delivery location must be served after its corresponding pickup location. The vehicle’s journey should start and end at a central depot, while the load carried by the vehicle should not exceed its capacity. In addition, each request must be served within a certain pre-determined time window (TW) interval. If the vehicle arrives at a location before its due service time, it must wait until the beginning of the specified period. A solution to the problem should assign requests to vehicles and generate a feasible route for each vehicle, such that the total service cost is minimized, and all problem constraints are adhered with. A formal problem definition can be found in [6]. Among the practical applications of the MVPDPTW are the e-commerce activities, the transport of raw materials from suppliers to factories, and the important dial-a-ride services, in which people instead of goods are transported.