The Study On P-Hub Center Problem In Fuzzy Environment

The p-hub center problem is to locate p hubs in a network and to allocate thenon-hub nodes to hubs such that the maximum travel time between any o-d (origin-destination) pair is minimized. In realistic p-hub center problem, the travel timeparameters usually present fuzziness, so we should employ fuzzy theory to deal withfuzziness in p-hub center problem.This thesis first presents a new class of fuzzy p-hub center problems based oncredibility criterion. In the case that the fuzzy travel times are mutivariate normalfuzzy vectors, we suggest an approximation approach (AA) to reformulating theoriginal problem as a mixed-integer problem with logic constraints. We design animproved hybrid particle swarm optimization (PSO) to solve the proposed model.Taking into account of the difcult of determining the possibility distributions in re-alistic p-hub center problem, we further develop a new class of p-hub center problemswith generalized value-at-risk (VaR) by a robust method. When the fuzzy traveltimes are independent type-2fuzzy variables, we derive the parametric possibilitydistributions of reduced fuzzy variables by applying reduction methods. In this case,we discuss its equivalent parametric mixed-integer programming problems and theefective solution methods. Finally, some numerical experiments are performed todemonstrate the modeling ideas and the efciency of the proposed solution methods.The major new results of this thesis contain the following four aspects:(i)Two classes of novel fuzzy p-hub center problems are proposed;(ii) Three kindsof asymmetric type-2fuzzy variables are defined, and the parametric possibilitydistributions of reduced fuzzy variables are presented;(iii) In the cases that traveltimes are mutivariate normal fuzzy vectors or asymmetric type-2fuzzy trapezoidalvariables, the proposed models can be turned into their deterministic equivalentones;(iv) An improved hybrid PSO algorithm is designcd to solve the proposedmodels, and some numerical examples are provided to illustrate the modeling ideasand the efectiveness of the proposed hybrid algorithm.