| Location-allocation problem is to study that how to locate the facility and how to allocate the products for the customers. In realistic location-allocation problems, the input parameters usually have fuzziness, so we should adopt a theoretical tool that deal with fuzziness to study location-allocation problem.Under this consideration, this thesis takes credibility theory as the theoreti-cal foundation to study location-allocation problem. First, this thesis presents two class of two-stage fuzzy location-allocation models-the expected value location-allocation model and the minimum-risk location-allocation model, and studies the properties of the second-stage problem. Considering customers demands are chara-terized by fuzzy variables that have infinite supports, we apply an approximation method to turn them into a discrete ones that taking finite values. Based on approx-imating method, we present the computational method that compute the expected value recourse function and the credibility recourse function. At last, we design a hybrid algorithm which combines approximation method and particle swarm opti-mization, and present one numerical example to illustrate the effectiveness of the designed algorithm.The main work of this thesis can be summarized as follows:(ⅰ) We present two new class of two-stage fuzzy location-allocation models-the expected value location-allocation model and the minimum-risk location-allocation model, and study the properties of the second-stage problem.(ⅱ) Based on the approximation method, we give the computational methods of the expected value recourse function and the credibility recourse function, and discussed the convergence about the objective function of the approximating problem to that of the original one.(ⅲ) We design a hybrid algorithm to solve the approximating problem, and give one numerical example to demonstrate the effectiveness of the hybrid algorithm.(ⅳ) We study the analytical properties of simple integer recourse fuzzy pro-gramming. |
PDF Full Text Request