Research On Logistics Distribution Problem And Algorithm With Uncertain Information

Logistics is a part of supply chain activities.It is a process of transporting goods from origin or distribution center to all demand locations.Logistics cost occupies a large proportion in the total operation cost of supply chain.Today,in an environment with increasingly fierce competition in the market economy,only compressing cost of goods has been unable to significantly improve the economic efficiency of enterprises,reducing the distribution cost has become the primary factor that related enterprises must consider.The efficient and low-cost implementation of logistics distribution is very important to the competition of supply chain.Therefore,the rational planning of the logistics distribution scheme is a very important link in the supply chain management.The vehicle routing problem in logistics distribution was first proposed by Dantzig and Ramser in 1959.It can be described as: A vehicle carrying a certain amount of goods starts from the distribution center,and then pass all the nodes with certain commodity demand to meet the needs of each distribution node.Finally,the vehicle returns to the distribution center.If the goods are insufficient,it can return to the center for replenishment in the midway.The goal of the problem is to plan a route with the shortest distance,that is,the order of distribution.In the actual supply chain activity,the market demand is usually uncertain random variable.Therefore,for a fixed driving route,it is uncertain that which nodes the vehicle need to return to replenish at and how many times to replenish.So,the traversing distance is random.This kind of problem is called Vehicle Routing Problem with Stochastic Demand(VRPSD),and its goal is to plan a route with minimum mathematical expectation of travel distance.In addition,distribution time is often uncertain,which is called the fuzzy time window vehicle routing problem(Vehicle Routing Problem with Soft Time Window,VRPSTW),and its goal is to plan a shortest distribution route satisfying the fuzzy time window constraint..From the point of view of the control of logistics distribution cost and improvement of logistics distribution efficiency,considering the randomness of customer demand,the loading capacity of delivery vehicles and distribution time,a mathematical model of vehicle routing problem under different strategies is established,and an optimization algorithm for finding optimal route is designed in this dissertation.The main research work is as follow:(1)Under the strategy of non-return early,two single vehicle route models with stochastic demand are established respectively with limited and unrestricted return times to replenish goods.The general mathematical expectation of vehicle driving distance is formulated,and the specific formulas for computing expectation are given respectively according to the two cases of demand following discrete distribution and continuous distribution.(2)Under the strategy of return early,the recurrence formulas of vehicle driving distance are given respectively according to the two cases of demand following discrete distribution and continuous distribution.Based on the hypothesis of demand following binomial distributions or Poisson distribution,the two strategies of return early and non-return early are compared.And the results show that the expectation of driving distance under the strategy of return early is smaller.(3)Based on the genetic algorithm to solve the TSP problem,the genetic algorithm for VRPSD is designed to minimize the mathematical expectation of driving distance,and the optimal distribution order is found.Finally,the feasibility and efficiency of the algorithm are verified by simulation examples,and the influence of the parameters of the operator on the convergence of the algorithm is analyzed.(4)Aiming at the vehicle routing problem with uncertain distribution time,taking minimizing the distance of distribution route and minimizing the number of vehicles as optimization objectives,a multi-objective programming model with fuzzy time window constraints is established,and a fuzzy linear programming method for solving the optimal distribution path is designed.Figure [13] table [4] reference [24]...