Research On Vehicle Routing Problem For Customer's Dynamic Demand Via Internet Of Things

In the face of COVID-19,which has been raging around the world and not yet been seen for a century,the Central Committee of the Communist Party of China upholding the supremacy of the people's life and made major strategic decisions timely,building a new development pattern of domestic and international double cycle and promoting the development of the smart cars and Internet of things(Io T),promoting the transformation and upgrading of the transportation industry through the "demand side reform" that links production,circulation and consumption.In the new environment of the economic growth and technological change,the logistics industry maintains a steady growth trend,at the same time,customer demand tends to be personalized,diversified and professional,therefore,it is a severe challenge to dynamic vehicle routing problem(DVRP).Under the Io T architecture,the dynamic information of customer demand is transmitted to the cloud in time.It is the key to reduce logistics operation cost,improve customer satisfaction and enterprise core competitiveness to effectively formulate dynamic scheduling scheme combined with change information and respond to customer demand change in a fast and low-cost way.However,the dynamic vehicle scheduling optimization process is a complex system engineering,and the factors involved have high dynamic correlation and complex nonlinear relationship.Therefore,building a systematic vehicle routing optimization method system for the customer's dynamic needs is one of the key scientific issues that need to be solved urgently.Based on this,the paper uses the front,middle,and end of logistics as application scenarios to systematically study the vehicle routing problem.Firstly,under the end delivery scenario,styding the one-stage static vehicle routing problem with time windows(VRPTW)considering the fixed customer needs;Secondly,under the front collection scenario,studing the one-stage dynamic vehicle routing problem with time window(DVRPTW);Thirdly,under the front and mid-end two-stage collection scenarios,studing the two-echelon dynamic vehicle routing problem with time window(2E-DVRPTW);Finally,on the basis of 2E-DVRPTW,the sharing economy is introduced to study the two-echelon dynamic shared VRPTW(2E-DSVRPTW)considering familiarity and codelivery.The main problems and research content of this paper are as follows:?In order to solve the issue of poor quality of vehicle routing for customers,a method for VRPTW based on improved parallel simulated annealing(I-PSA)is studied.Firstly,a mathematical model with vehicle using and driving cost as the primary and secondary optimization objectives is constructed.Secondly,the I-PSA algorithm for VRPTW is designed based on master-slave parallel mode by introducing multi Markov chain.Then,the validity of the proposed method is verified by using the Solomon data set;The literature comparison method and double sample hypothesis test method be used,and the differences of I-PSA,particle swarm optimization(PSO),ant colony tabu hybrid(ACO tabu)and ant colony optimization(ACO)are compared and analyzed;The sensitivity of different parameters to the optimization results is analyzed by using the experimental method.The results show that the proposed method has certain advantages over the existing literature optimal solution.Taking R101 as an example,the optimization rate reaches 1.31%,which has a guiding role for enterprises to realize the dual optimization of cost and customer satisfaction.?In order to solve the issue that the dynamic demand of customers disturbs and destroys the VRPTW optimization results,the DVRPTW problem and its solution are studied.Firstly,the idea of Bayesian condition,time slice and dynamic degree is introduced to optimize the path of the next time slice based on the premise that the internal path of the previous time slice is effective.Secondly,a tabu search(TS)algorithm is designed to solve DVRPTW.Then,the DVRPTW generation method and data given by Kilby are used to analyze the effectiveness of the proposed method;Finally,an enterprise is taken as an example to verify the proposed method,and the sensitivity of dynamic degree to optimization objective is analyzed by using the experiment method.The results show that there is a gradient positive correlation between the dynamic degree and the distance and use cost of vehicles.When the dynamic degree increases in a continuous interval,the use cost of vehicles presents a discontinuous gradient jump increasing feature.The proposed method is of great significance for enterprises to reduce the cost of responding to customers' dynamic needs and improve the efficiency of delviery.?In order to solve the problem of low delivery efficiency of staff and low loading rate of vehicles,a solution strategy of 2E-DVRPTW considering regional familiarity and load balancing is designed.Firstly,the K-means-based region partition strategy and the load capacity balancing method of the transfer station are proposed,and the method of mapping the continuous two-layer main network optimization to two related networks is proposed,which is integrated into the solving process of 2E-DVRPTW.Secondly,based on the data set given by Perboli,the performance of the model and algorithm is tested by using literature comparison method.Compared with the optimal solution in the literature,the optimization rate of the proposed method is up to 15.76%.Then,taking an enterprise as an example,the optimization results of four scenarios are compared and discussed,including not considering regional familiarity,considering regional familiarity,considering regional familiarity and load balancing,and considering regional familiarity and dynamic degree.The purpose of the design method is to provide scientific guidance for enterprises to carry out double-layer dynamic delviery tasks.?In order to solve the problem of double-layer dynamic vehicle scheduling with shared vehicle involvement,based on the research results of 2E-DVRPTW,the modeling and solving method of 2E-DSVRPTW are further studied.Firstly,with the objective of minimizing the distance between self owned vehicles,shared vehicle service nodes and transfer stations,the optimal number and location of transfer stations are determined based on K-means method;Secondly,with transfer stations as the dividing point between self owned vehicles and shared vehicles,the network composed of centralized centers and transfer stations and the network composed of transfer stations and customers are dynamically optimized by using 2E-DVRPTW method.Then,the validity of the model and algorithm is analyzed and verified by using the international general augerat data set.Then,the control variable method and simulation comparative experiment method are used to analyze the sensitivity of algorithm parameters to the optimization results;Then,the literature comparison method and double sample hypothesis test method are used to analyze the differences between the designed model and other model algorithms.Finally,taking an enterprise as a case,the preliminary application of the designed 2E-DSVRPTW method is verified,and the results are compared with those of different schemes.The purpose of the proposed method is to help enterprises make scientific decisions on 2EDSVRPTW considering the dynamic needs of customers.In conclusion,the research results of this paper are expected to improve the timeliness,accuracy and scientificity of vehicle response to customers' dynamic demand changes under Io T environment,and enrich the basic theoretical system of vehicle routing modeling and dynamic optimization in Io T scenarios.