Fuzzy Raw Materials Allocation Problem

AbstractBased on credibility theory, this paper presents a new raw materials allocation (RMA) model in which the demands of products and the prices of surplus products are supposed to fuzzy variables, the involved restrictions mainly include the inventory of each raw material and the production capacity of company. Since the fuzzy variables in the RMA model are usually continuous variables and the exact analytical expression of objective function cannot be obtained, traditional optimization algorithms cannot be used to solve it. Therefore, we firstly apply an approximation approach (AA) to turn the original infinite-dimensional optimization problem into an approximating finite-dimensional one, then design a hybrid algorithm which combines neural network (NN) and particle swarm optimization (PSO) to solve the approximating model. A given numerical example illustrates that the designed algorithm works well