Research Of Two-stage Fuzzy Material Procurement Planning Problem

Material procurement planning (MPP) deals with the problem that purchasing the right quantity of material from the right supplier at the right time, the manufac-turing companies can reduce the material procurement costs via a reasonable MPP model. However, in the real material procurement process, many factors have the fuzzy uncertainty, so we need a kind of theory to handle the fuzziness. In this thesis, we employ credibility theory to study the MPP problem in a fuzzy environment.Based on different modeling ideas, we present two new kinds of two-stage fuzzy material procurement planning models:the expected value material procurement planning model is to minimize the expected material procurement cost; while the minimum-risk material procurement planning model is to minimize the material pro-curement risk with a given capital investment. We use approximation approach (AA) to turn the two kinds of infinite-dimensional optimization models into approximat-ing finite-dimensional problems, and the convergence about the objective function of the approximating MPP model to that of the original MPP one is discussed, re-spectively. In addition, we design an AA-based particle swarm optimization (PSO) to solve the proposed two-stage fuzzy MPP models and some numerical examples are given to illustrate the effectiveness of the designed algorithm.The main work of this thesis can be summarized as follows:(1) Based on different modeling ideas, we present two new kinds of two-stage fuzzy material procurement planning models:the expected value material procure-ment planning model and the minimum-risk material procurement planning model.(2) We give the computational methods of the expected value objective function and the credibility objective function in the two kinds of fuzzy MPP models, and discuss the convergence about the objective function of the approximating material procurement planning model to that of the original one, respectively.(3) We design an AA-based PSO algorithm, and some numerical examples are given to illustrate the effectiveness of the designed algorithm.