Research On The Balancing And Scheduling Problem Of Sequential Multi-Stage Automated Flexible Production Line With Parallel Machining Cells

The diversity and individualization of the customer requirements and the fierce market competition have forced the manufacturing company to continuously improve the flexibility and intelligence of their manufacturing system.How to design a flexible manufacturing system that satisfies the requirements of the company and achieve an efficient and balanced output is one of the key issues in the current development of the digital factory and intelligent manufacturing.As the crucial component of the flexible manufacturing system,rationally configuration of the manufacturing resources for the flexible production line in order to balance and optimize the line,maximization of the potential capacity of the line and minimization of the production cost of unit production,have become a hot research topic widely concerned by both academia and industry.The thesis has focused on the requirement of the flexible production line in a mobile terminal processing enterprise,and research on the flexible production line balancing and configuration problem.On the basis of two basic line balancing problems,the flexible production line balancing problem under the condition of machine failure and uncertain maintenance time,the flexible production line balancing problem under the multi-objective with uncertain processing time and the lot-sizing and scheduling problem for the flexible manufacturing system with multiple flexible production lines are deeply studied.The characteristics of the flexible production line,the mathematical modeling and optimization algorithms have been investigated,including the following aspects:(1)A new type of flexible production line with sequential manufacturing stages consisting multiple machining cells is studied,the structure and features of the flexible production line are studied,and the cycle time relations between the stages,cells and the line are indicated.In the flexible production line balancing problem type I,the precedence constraints,exclusion constraints and inclusion constraint between the operations are considered,and a mathematic model is proposed with minimizing the line configuration cost in required line cycle time.The computational complexity of the problem is proved to be a NP-hard problem.The branch and bound based heuristic is proposed for the problem and several experiments based on the real situation are made to prove the efficiency of the algorithm.And the final results have proved that the proposed branch and bound based heuristic is efficient to solve the flexible production line balancing problem in reasonable time.(2)The line balancing problem type II based on the proposed new mode of flexible production line is studied.A mixed integer programming is made with the objective by minimizing the line cycle time in the situation with known number of machines and robots.Moreover,a novel method named set partitioning based heuristic(SPH)is proposed to solve this new flexible production line balancing problem to minimize the line cycle time and the performance is compared with both exact algorithm(branch and bound algorithm)and random search algorithm(tabu search).A set of benchmark instances based on different size of problems against different system parameters is made.Furthermore,sensitivity analysis of the system parameter is performed to know,how the number of machines and processing time can influence the cycle time and the utilization of line.Computational experiments are performed to show the performance of the proposed method SPH against other methods and the results indicate that SPH performs best among all the test methods in terms of solution quality and computation on all the proposed benchmark instances.(3)Focus on the dual-uncertainties including uncertain occurrence of machine failure and uncertain maintenance time in the flexible production line,the thesis research on the corresponding line balancing and configuration problem,and established the chanceconstrained programming modeling,and proposed a solution method named decomposition heuristic,which efficiently decompose the chance-constrained programming and linearize the chance constraints.By inducing the stochastic simulation method and experimental examples in different scale,the results between the two methods have the consistency in the quality,and the decomposition heuristic has better computational speed efficiency.In addition,the sensitivity analysis of reliability by adding machines to the line has been investigated.Moreover,the proposed mathematic model and the algorithm have been used to solve the selection of the machine supplier as a preliminary application.(4)The flexible production line balancing problem with the aim to minimize line cycle time,minimize smoothness index and maximize the reliability of desired line cycle time simultaneously.The considered multi-objective problem with uncertain processing times is solved with robustness approach.Three new multi-objective robust dominance criteria are defined based on scenario planning approach.Moreover,a novel heuristic based on branch and bound algorithm(HBB)and heuristic based on artificial bee colony algorithm(HABC)are developed to get multi-objective robust solutions of the current problems based on the proposed multi-objective robust dominance criteria.Experiments are performed to solve different size of machining line balancing problems.The performance of HABC is compared with the HBB against different problems with different scenarios based on proposed multi-objective robust dominance criteria.Computational results indicate that HABC outperforms HBB according to the all defined three multi-objective robust dominance criteria and HBB is quicker than HABC in CPU time.Overall results indicate that both HBB and HABC are significant to give multi-objective robust solutions of the considered flexible production line balancing problem with multiple conflicting objectives.(5)In flexible manufacturing system with multiple flexible production lines,continuous arrival of customer orders with different size of demand and due dates,makes it hard for company to make allocation of demand of customer orders on multiple production lines in each planning horizon.Dynamic lot sizing and scheduling on multiple lines to meet the customer due dates is significant in multi lines production environment.Therefore,current research investigates dynamic lot-sizing and scheduling problem in multiple flexible production lines environment considering mixed model products.In addition,uncertainty in demand and machine failure are considered.Moreover,sequence dependent setup time between mixed model product lots is also included during making the schedule.A mixed integer programming model is proposed for the considered problem with an aim to maximize the probability of completion of product models from different customer orders on multiple lines in their due dates in each planning horizon.A constructive heuristic for lot sizing and scheduling problem(CHLP)is proposed to solve the current problem.The proposed CHLP heuristic involves the steps to distribute different demand of customer orders among multiple lines and schedule them considering balancing of makespan between the lines.Different sizes of test problems are created and performance of the proposed heuristic is measured with famous heuristics in literature,based on the created test problems instances.Results indicate that CHLP gives better results in terms of quality of results for most of the problem instances as compared to the considered famous heuristics in literature.(6)The above theories and research results are integrated into the advanced planning and scheduling system named Xplanner,which is developed by our research group.In view of the optimization algorithm proposed in the paper,a sub-module named "line configuration and optimization system" is developed and the optimization algorithm library is enriched.Finally,a real case problem of company H under the background of a new mode of an intelligent manufacturing has been solved efficiently,which also fully clarifies the research and application value of this paper.The results of this paper enrich the research on the balancing and optimization of automated flexible production line to a certain extent.The new-type automated flexible production line and the key technology proposed in this paper not only have a theoretical breakthrough,but also can effectively improve the utilization of enterprise’s resource utilization and ability in decision-making management.