Research On Vehicle Routing Optimization With Fuzzy Demand Under Time-varying Conditions

With the development of global economy,the current market environment is also undergoing profound changes.As the third profit source,modern logistics is getting more and more enterprises' attention.As a link connecting consumers and producers,logistics plays an increasingly prominent role.Logistics transportation is an important part of modern logistics operation.According to the statistics of China Federation of logistics and purchasing,the logistics circulation rate has increased significantly in recent years,and transportation cost accounts for the vast majority of modern logistics cost.Therefore,it is one of the hot issues to reduce the logistics cost by optimizing the vehicle route of logistics distribution.In the research of traditional vehicle routing problem,the solution algorithm and solution strategy are the two main research contents of researchers at home and abroad.With the development of modern logistics,the current research on vehicle routing problem is more and more close to the reality of logistics distribution activities,and the traditional vehicle routing problem model can not accurately reflect and solve the actual problems.Therefore,the current research on vehicle routing problem is mostly the development of traditional vehicle routing problem.For example,there are vehicle routing problems with uncertain customer demand.In the research of this kind of problems,many researchers comprehensively consider the constraints of customer fuzzy demand and time window,but only consider the situation of constant vehicle speed,ignoring the impact of weather changes,peak hours,emergencies and other factors on the traffic conditions,resulting in the fuzzy demand base'd on constant speed The model of vehicle routing problem is no longer applicable.There is also a kind of time-dependent vehicle routing problem,many of which are based on the customer's demand is known and determined,without considering the fuzziness and uncertainty of the customer's demand in real life.The main content of this paper is the time-dependent vehicle routing problem with fuzzy customer requirements and time window constraints.The influences of time-varying road network speed,fuzzy customer demand and time window are considered.To solve the TDVRP problem with fuzzy customer requirements and time window constraints the model is built on the idea of pre-optimization and rescheduling.In the pre-optimization stage,the fuzzy opportunity constrained optimization model is constructed according to the credibility theory to deal with the fuzzy demands of customer points.Besides,Figliozzi speed time-dependent function was used to represent the driving speed of vehicles according to the traffic conditions of roads in different time periods.To solve the model an adaptive large neighborhood search algorithm(ALNS)is designed.In the rescheduling stage,stochastic simulation algorithm is used to simulate the real demand of customer points and point rescheduling strategy is adopted to adjust the pre-optimization scheme.The effectiveness of the model and algorithm is verified by an improved Solomon example.The experimental results show that in the TDVRP problem with fuzzy customer requirements and time window constraints,the actual cost of the optimal pre-optimization scheme is not necessarily the lowest.In the pre-optimization stage,with the increase of the decision maker's preference value,that is,the more conservative the decision maker tends to be,the higher the distribution cost will be;in the rescheduling stage,the more conservative the decision maker tends to be,the lower the extra cost will be.The research results can enrich the related research of TDVRP problem and provide theoretical basis for the optimization decision of realistic distribution scheme.