Research On The Flexible Resources Scheduling Problem And Optimization Algorithm For Manufacturing Systems

With the constant development of advanced manufacturing technologies and modes, and the diversification demand of market, as well as the application of industry informatization, manufacturing enterprises are facing tremendous challenges and opportunities. Manufacturing environment is moving towards more diversified resources, more common share, more frequent market changes and more fierce competitions. Optimizing the dispatching process of flexible resources in varied manufacturing systems is beneficial to the full utilization of resource share and the maximization of enterprise benefits. But as the complexities of manufacturing system and amount of flexible resources are expanding, the complexity and difficulty of optimal scheduling is increasing rapidly at the same time.This thesis is mainly focusing on the two main problems in Flexible Resources Scheduling. One is optimal scheduling of flexible resources in one single enterprise, and then further studies the model and optimizing algorithm of flexible resources scheduling in multi-enterprise web-based manufacturing system. The key point of the research is algorithm optimization and simulation.Firstly, manufacturing system and its flexible resources are introduced. Then this part analyzed and compared the resource scheduling problem in one single enterprise system and multi-enterprise web-based manufacturing system so as to figure out the commonalities and differences of optimal scheduling for multi-granularity flexible resources in systems of different scales.The second part is mainly focusing on the further study of most widely discussed Flexible Job-shop Scheduling Problem in one single enterprise. Through description, analysis and model of FJSP, the basic technical route based on genetic algorithm and two main directions of optimization, aiming at the optimization of machine selection initialization in genetic algorithm and neighborhood search technology in local search respectively were proposed. Through the optimization, a hierarchical algorithm based on makespan limit and a hybrid genetic algorithm with variable neighborhood search based on expanded idle time are designed. The feasibility and effectiveness of the algorithm is verified through tests on international standard benchmarks and practical manufacturing cases.Last part of this thesis is focusing on flexible networking resources scheduling problem for multi-enterprise in web-based manufacturing system, analyzing the expression of tasks and resources. Logical manufacturing units and physical manufacturing units are used for unified scheduling of manufacturing tasks and resources of different granularities and geographic locations. Multi-objective optimal scheduling model is proposed based on two evaluation indicators of makespan and manufacturing costs, then get solved with optimized hybrid genetic algorithm. Feasibility of the model and algorithm is verified with tests on simulation examples.