A biased random key genetic algorithm applied to the VRPTW with skill requirements and synchronization constraints

ABSTRACTWe applied a Biased Random Key Genetic Algorithm (BRKGA) to solve the Vehicle Routing Problem with Time Windows and Synchronization Constraints. Additionally, both vehicles and clients are skilled, and each client can require up to two distinct skills to be serviced. On double-skilled clients, the operations of each skill should be performed by different vehicles, either simultaneously or respecting a precedence order. Those requirements introduce nonlinearities on the problem, in the sense that a small change on a single route potentially impacts all the other routes of the solution, making it hard to define an effective local search procedure. To circumvent this difficulty, we approached the problem using a genetic algorithm that evolves the sequence in which the services are inserted into the routes. We assessed the performance of our solution method using instances from the literature of the home health care problem. The genetic algorithm outperformed the previous best-known solutions found by a fix-and-optimize matheuristic by up to 25%, using less than half of computational times reported previously. The BRKGA demonstrated to be able to perform well both in exploration and exploitation in the solution space of the problem.