Study On Vehicle Routing Problems Considering Time-Dependent Road Network Under The Dual-Carbon Target

"Carbon peaking" and "Carbon neutrality" are not only our national development strategy,but also a solemn commitment made by the Chinese government to the people all over the world.The core of achieving the "Dual-carbon target" is to build a green,lowcarbon and circular economic development system,improve the utilization rate of energy and resources,cut down the proportion of fossil energy consumption,and reduce the level of carbon dioxide emissions.With the rapid development of urbanization and motorization,the population has gathered in cities and the number of auto vehicles has increased year by year,which leads to the frequent occurrence of traffic congestion.Traffic jams not only push up the cost of urban logistics,but also reduce the delivery timeliness of orders.The consequence is not conducive to the healthy development of urban logistics.China’s annual express delivery volume has exceeded 100 billion for two consecutive years since 2020,such a large volume of good flow needs to be delivered within a limited time,which brings great challenges to the distribution network within cities.In addition,the emission resulting from a large number of conventional vehicles is of great impact on the environmental quality of the central urban area,and the carbon emissions and environmental pollution caused by the trucks cannot be overlooked.The above concerns affect the logistics cost and living environment quality of residents directly,and even against the development direction of green logistics.Therefore,optimizing urban logistics operation under the traffic congestion environment can not only help logistics enterprises reduce costs and increase efficiency,but also lower urban pollutant emissions and achieve the "Dual-carbon target".Although existing studies have made a lot of achievements in the modeling and algorithm design of low-carbon vehicle routing problems,there is a lack of research on the estimation of time-dependent road network parameters in the target region and how to implement the planed distribution scheme,and there is more room for improvement in the solution algorithms for bi-objective or multi-objective problems.On this basis,the low-carbon vehicle routing problem is taken as the research object,and the construction of route optimization models,the design of solution algorithms and case analysis verification are carried out according to different application scenarios of urban logistics.This research provides a reference for the practical problems faced by the decision makers of logistics enterprises,such as time-dependent road network parameters estimation,distribution network construction or optimization and distribution scheme fulfillment.The main contents and achievements of this study are as follows:(1)Time-dependent road network research based on deep learning methodsThe real-time traffic congestion index of the target area can be crawled by using the application program interface provided by online map platforms.A deep learning prediction model based on stacked recurrent neural network is constructed to predict the traffic performance index in the future distribution process.The basic traffic information between distribution network nodes is requested from online map platforms under the condition of smooth traffic flows.Finally,combining the predicted time series of traffic performance index with basic traffic information,the time-dependent network cost matrix is estimated by using the segmented time-dependent travel duration calculation method.These information will be directly applied to the subsequent models and algorithms of low-carbon vehicle routing problems.(2)Research on the low-carbon vehicle routing problem under time-dependent road networkUnder the time-dependent road network,a vehicle routing problem model considering economic objective and environmental objective is established,and a heuristic algorithm based on the clustering method is designed.The model and algorithm are suitable for route optimization of urban logistics in traffic congestion environment.At the same time,based on the multi-attribute decision making theory,a sustainable logistics scheme evaluation framework including the principal component analysis and simple additive weighting is proposed to provide objective and quantitative analysis results for decision makers to choose the most satisfactory logistics scheme.(3)Research on the multi-depot low-carbon vehicle routing problem under timedependent road networkA multi-depot low-carbon vehicle routing problem model and a corresponding goalprogramming model are presented considering the time-dependent road networks.Under the framework of traditional genetic algorithm,a hybrid genetic algorithm with intensification procedures is designed by using the three-level linked list structure.Combined with lexicographical multi-objective optimization theory,a new selfconstrained hybrid genetic algorithm is proposed to solve the bi-objective model.Combined with the road network of target areas,an online route update strategy is developed to guide remote vehicles against the real-time changes in traffic flow and to record detailed distribution paths.Finally,the distribution scheme is displayed on the actual road network layer by using the route fitting technology,which is beneficial to decision-makers to visually observe the distribution plan and facilitate subsequent statistics and analysis.(4)Research on the multi-objective two-echelon low-carbon vehicle routing problemBased on the two-echelon vehicle routing problem model,a multi-supplier twoechelon vehicle routing problem optimization model considering heterogeneous fleet is presented.Multiple objectives are defined to represent the concerns of logistics enterprises,government departments,customers or the general public.According to the characteristics of the problem,a parallel multi-objective evolutionary algorithm is designed to obtain Pareto frontier solutions.Finally,the carbon emission evaluation framework of fuel vehicles and electric vehicles considering time-dependent road network is designed,and the decision-making method of hybrid fleet configuration under different carbon emission regulations is introduced to facilitate the realization of the "Dual-carbon" target.