The Genetic Algorithm Scheduling of Vehicles from Multiple Depots to a Number of Delivery

Abstract:- We develop a new heuristic based on the genetic algorithm (GA) that is used to find high quality solutions to the non-fixed destination Multiple Depot Capacitated Vehicle Routing Problem (MDCVRP) in a reasonable amount of computing time. The definition of the problem with more adherence to reality was proposed. More advanced issues such as solution representation and decoding are developed that enable simultaneous routing of all vehicles. The problem constraints (quantity of goods available at depot, vehicle capacity, maximum route length and maximum number of customers visited in one route) are relaxed using the penalty terms whose parameter values are adjusted during the GA run. Six well-known crossov r and three mutation operators developed for the sequencing type representations were compared on fifteen test problems each with different geographical and temporal characteristics. Computational experience one real instance of MDCVRP with 3 depots and 248 customers is discussed that should be didactic illustration of the characteristics and originality of the presented approach