Research On Location Problem Of Logistics Center Under Uncertainty

Logistics center location is a process in which one or more distribution centers are located in an economic zone having a certain number supply and demand points. Traditional location model of logistics center usually takes some key factors in logistics system as known constants, such as holding cost, transportation cost, demand and transportation time etc., and then one or more locations are sited according to requirements. In practice, some factors may be uncertain. For example, demand may be fluctuated; transportation time may be restrained by traffic conditions. These possibilities form the uncertain conditions for logistics center location. So we should take these uncertain factors into consideration when programming complex logistics system.The existence of uncertain factors could make the decision-making system of logistics center location contain stochastic and fuzzy parameters. But uncertainty is not purely decided by the single factor, sometimes it is the combination of the two or more factors. This paper revises the definition and hypothesis of the classical logistics center location model—the Kuehn-Hamburger model, and presents a basic and expanded time chance constrains model under uncertainty, in which the demand is stochastic and transportation time is fuzzy. This model is general for handling uncertain factors.For seeking the solution of uncertain model, this dissertation discusses the theory of chance constrains programming with fuzzy and stochastic parameters, fuzzy simulation, and stochastic simulation, and the presents a way of handing fuzzy and stochastic parameters of the model, which uses fuzzy stochastic simulation, lastly designs a genetic algorithm based on fuzzy stochastic simulation, and the genetic operation, such as coding, mutation, selection and crossover etc.Finally, the nearly optimum solution is calculated by using simulation data and the genetic algorithm based on fuzzy stochastic simulation, which illustrate the feasibility and validity of this model and algorithm.