The Vehicle Routing Problems In The Two-Echelon Logistics Transportation System

Logistics and transportation industry is one of the pillar industries that have a bearing on the country's economic development and social progress.At present,the need for logistics and transportation in our country increases year by year,but the development is still in its infancy.As an industry with low profitability,cost management is very important.Multi-echelon logistics mode is emerging in recent years.The multi-echelon division of the transportation network can ease the problems caused by single-echelon logistics such as traffic congestion and environmental pollution,reduce the total logistics costs and promote the sustainable development of the logistics industry.As the most basic mode of multi-echelon logistics,two-echelon logistics transportation has been extensively researched and explored.Based on this,the paper takes into account the characteristics of simultaneous customer delivery and pickup demands,the uncertainty of customer demands and stochastic travel times,and studies the transportation problems of two-echelon logistics.In the basic two-echelon logistics and transportation,the network is divided into two echelons: the first-level vehicles with large capacity depart from the main depot,serve the satellites and then return to the depot;second-level vehicles with small capacity start from each satellite,meet the needs of one or more customers,and then return to their respective satellites.In the scenario of two-echelon logistics and transportation with simultaneous delivery and pickup demands,the vehicles must satisfy both demands of each customer at the same time.Therefore,the entire transportation process is divided into three phases.With regard to the case with uncertain customer demands,the capacity of the vehicle may not be able to meet the total demands of the pre-set service customers during the actual transportation process.Therefore,in actual operations,uncertainties in customer demands can result in route failures when vehicles operate the planned routes and need recourse actions.In the two-echelon logistics scenario with stochastic travel times,the penalty cost of travel times exceeding the maximum vehicle working time is introduced into the total cost function.Due to the randomness of travel times,the penalty cost that vehicles may cause in actual transportation is also uncertain.For the case with simultaneous deliveries and pickups,a variable neighborhood search based algorithm(VNS)is designed where feasible and infeasible region are both searched.In order to increase the solving speed,the annealing mechanism is used.To solve the problem with stochastic demands,a simulation based tabu search(STS)is developed and Monte Carlo simulation process is embedded into the algorithm to evaluate the neighborhood solutions.As for the case with stochastic travel times,the STS algorithm is modified and improved for solving the specific new problem better.Afterwards,based on the standard instances in the existing literature,new instances are generated for each problem.Parameter test experiments are carried out for determining the parameter of simulation times in the STS.Computational experiments on instances of different sizes compare the proposed algorithms and their simplified versions,respectively.The results prove that the proposed approaches are adaptable and efficient.Meanwhile,the developed VNS algorithm is adjusted to solve the benchmark instances of standard two-echelon vehicle routing problem.After comparisons with the best known solutions on these instances,the VNS algorithm is proved to be stable and of high quality.