Research On Model And Algorithm Of Low Fuel Consumption, Multi-depot, Multi-model Vehicle Routing Problem

In order to expand the business of enterprises and meet the needs of customers in time with the increasing development of logistics and distribution,many enterprises have set up multiple distribution centers and provided a variety of types of vehicle for different types of goods.At the same time,to enact a bill to protect the environment are increasingly strict by governments,and low-energy vehicle delivery,which takes into account factors such as fuel consumption and carbon emissions,has become a priority.In addition,the increasingly fierce market competition forces enterprises to focus on reducing logistics distribution costs and improving customer satisfaction.In this paper,it is of great practical significance to study the low-fuel-consumption multidepot heterogeneous-fleet vehicle routing problem with time windows(LMHFVPR?TW).In this paper,the vehicle routing problem(LMHFVPR?TW)with low fuel consumption and multi-depot heterogeneous-fleet with time windows and the intelligent algorithm to solve the problem are studied as follows:Firstly,in the decomposition phase of the problem,since the considered problem is a complex one with strong constraints,large scale and NP-hardness,In order to control the scale of problem and reasonably guide the algorithm to search in the high-quality solution region,one kinds of clustering method based on K-means strategies are designed to reasonably decompose it into a series of subproblems(i.e.,the low-fuelconsumption vehicle routing problems with time windows,LVRP?TW)by utilizing the problem characteristics.Secondly,In the solving phase of the problem,two-stage optimization is adopted.In the first stage,order optimization is carried out for each LVRP?TW.To get a good solution for each subproblem LVRP?TW,an enhanced ant colony optimization(EACO)to solve the decomposed subproblems(LVRP?TW)for obtaining the solution of the original problem is proposed.Not only a control factor of pheromone decay parameter in EACO is added to adjust the pheromone decay parameter dynamically so as to control the volatilization of pheromone effectively and improve the global search ability of ACO,but also an improved two-phase variable neighborhood search(ITVNS)is designed based on various variable neighborhood operations to enhance its the local search ability.In the second stage,vehicle speed is optimized based on LVRP?TWs optimal solution obtained in the first stage to further reduce the fuel consumption and the penalty of time windows of each sub-problem LVRP?TW,this paper proposes a speed optimization argrithm(SOA)based on improved hybrid differential evolution(IHDE)to optimize the vehicle speed of each sub-problem LVRP?TW.The method of changing the adaptive scaling factor by establishing the relation between the evaluation function and the number of iterations can suppress the problem of premature convergence of DE to some extent.Futhermore,simulated annealing algorithm(simulated annealing,SA)is added on the choice phase of algorithm in order to enhance the global search ability.Through the adoption of IHDE optimizing the speed of the vehicles can make LMHFVRP?TW further improve the quality of the solution.Finally,simulation experiments and comparisons on instances with different scales demonstrate the effectiveness of the proposed EACO?CD?SOA.