Some Researches On Fuzzy Flowshop Scheduling Problems Based On Hybrid Intelligent Optimization Algorithm

Production scheduling is the core and key technology of production organization and management. Making appropriate and optimized scheduling policy can improve the comprehensive management level of enterprises and bring economic and social benefits. Scheduling problems are usually complex optimization problems with multi-constraints, multi-objectives and various uncertain factors. It is significant but more difficult to solve the scheduling problems with constraints and uncertainties which are closer to the actual production environment. Several types of complex flowshop scheduling problems with different constraints and uncertain processing time are studied in this dissertation. Some proper fuzzy scheduling models are established with the tool of fuzzy programming theory. Since the traditional optimization methods can not solve the complicated scheduling problems, some hybrid intelligent optimization algorithms are proposed for specific issues. The main contributions of this dissertation can be summarized as follows:(1) The earliness/tardiness flowshop scheduling problem with uncertain processing time and distinct due window is concerned. An effective scatter search based particle swarm optimization (SSPSO) algorithm is proposed to minimize the sum of total earliness and tardiness penalties. The triangular fuzzy number is applied to describe the imprecise processing time of products and a fuzzy scheduling model is established by employing the algorithm of maximizing the membership function of middle value. The proposed SSPSO algorithm incorporates scatter search (SS) into the frame of particle swarm optimization (PSO) and gives full play to their characteristics of fast convergence and high diversity. In addition, differential evolution (DE) scheme and critical conditions are applied to improve the performance. The simulation results indicate the superiority of SSPSO to solve the earliness/tardiness flowshop scheduling with uncertainty.(2) Time-constrained intermediate storage flowshop scheduling with uncertain processing time is concerned. An approach for ranking fuzzy numbers is used to estimate the value and uncertainty of the makespan and employed to establish the fuzzy scheduling model. An improved particle swarm optimization with estimation of distribution algorithm (IPSO-EDA) is proposed. The IPSO-EDA incorporates the global statistical information collected from personal best solutions of all particles into the particle swarm optimization (PSO), and therefore each particle has comprehensive search ability. Meanwhile, the NEH-based initialization and local search are introduced to construct good initial solutions and enhance the local exploitation, respectively. In addition, the influence of parameter settings of the IPSO-EDA is investigated based on the method of factorial design. The simulation results indicate the superiority of the proposed scheduling algorithm.(3) An improved estimation of distribution algorithm (IEDA) is proposed to solve the fuzzy hybrid flowshop scheduling with parallel machines and uncertain processing time. The vector representation and an approach for ranking fuzzy numbers are used to establish the scheduling model. The proposed IEDA adopts the vector representation based coding method, consistently. The strategy of destruction and construction is applied to generate the initial population and improve the population diversity. Moreover, the destruction and construction strategy based variable neighborhood search is applied to the best solution to escape from local optimum. In addition, an orthogonal experiment design is performed to set the parameters of IEDA. The simulation results demonstrate that the proposed IEDA scheduling algorithm is effective and efficient.