Research On Flexible Job-shop Scheduling Under Uncertain Information Environment

Production scheduling plays an important role in improving efficiency and reducing costs. One of its key technologies is effectively modeling and its corresponding optimization algorithms. Nowadays, most researches concentrate on the optimization algorithms for classical scheduling problems without considering the uncertainties typically existing in the actual job shops, which leads to the difficulty to apply the algorithms in practical shop scheduling. In this context, how to build a more shop-approaching model and to design its corresponding optimization algorithm are very meaningful.Firstly, a Flexible Job-shop Scheduling model under uncertain information environment is proposed in this thesis. On the basis of describing the uncertain information existing in the actual job shops by several stochastic variables, a stochastic multi-objectives and multi-priorities programming model for job-shop scheduling is given out. In this model, time, cost and effectiveness of schedule are taken as three basic objectives of production scheduling. The scheduling constraints include processes of parts and the believable degrees that the delivery deadlines of different types of work pieces are satisfied.The structure of the above-mentioned scheduling model is too complex to solve by traditional precise algorithms. A strategy of hybrid intelligent algorithm combining Stochastic Simulation, Neural Network with Genetic Algorithm is proposed. It has two main steps. Firstly, a scheduling prediction model is built based on the approximation of uncertain scheduling jointly using Stochastic Simulation and Neural Network. Secondly, the approximated scheduling model is solved by Genetic Algorithm. According to this strategy, a specific algorithm is designed and implemented. A case study is given to illustrate the feasibility of this model and this solving approach.On the basis of theories mentioned above, a prototype system of hybrid intelligent optimization is developed. A good scheduling performance of this system is validated with SIMUL8 simulation tool in the practical shop.