Study On Green Time-Dependent Vehicle Routing Problem With Path Flexibility(GTDVRP-PF)

Traffic congestion worsens with urbanization.For urban logistics companies,congestion affects not only the accessibility and timeliness,but also the fuel consumption cost.It is one of the key factors affecting its operating costs.On the other hand,fuel consumption is related to carbon emissions.According to relevant statistics,the pollution of logistics accounts for 40%-50% of the total pollution.Therefore,the vehicle routing problem(VRP)with the minimum operating cost is to reduce the carbon emissions and reduce the operating costs of the logistics company.VRP is divided into two layers.One is the virtual customer layer,where the nodes represent customer points and the arcs represent the routes.The other is the real geographic layer,where the nodes represent the intersections and the arcs represent the roads.An arc on the customer layer represents multiple paths on the geographic layer based on different departure times.VRP is divided into a vehicle-customer allocation and routing problem at the customer layer,and a path selection problem at the geographic layer.Path Flexibility(PF)is introduced to achieve comprehensive decision-making at both layers.Based on the discretized departure time,path flexibility pre-calculates the minimum travel time path sets and embeds relevant information into the model.According to the Green VRP(GVRP),a compound function of operating costs is constructed,consisting of fuel consumption cost and vehicle depreciation cost.The Comprehensive Modal Emission Model(CMEM)is chosen for the fuel consumption function,considering the effects of speed,load and distance.According to the time-dependent VRP(TDVRP),the assumption,the travel speed is a step function of the time of the day,is chosen to construct the time-dependent road network,which better describes the congestion degree.A modified Dijkstra algorithm is used to preprocess the alternative path sets.The constructed GTDVRP-PF is more reasonable for urban logistics.It meets both economic and environmental benefits.The model is solved by genetic algorithm.The classic Sioux-Falls road network is used for numerical tests.The results are compared with the results of the minimum time and minimum distance to verify the effectiveness of the model and algorithm.There is some further analysis of the influencing factors such as waiting time,path flexibility and vehicle load.