The Study Of Cyclic Scheduling Method In Automated Production Systems With Flexible Processing Times

Automated production systems,which utilize industrial robot as the equipment to transport the materials,are widely employed in the semiconductor manufacturing,plating processing of printed circuit boards(PCBs),and other manufacturing industries.To manage the production of mass production effectively,cyclic scheduling production mode is usually employed in automatic production system.That is,the robot performs a fixed sequence of transport operations in each production cycle.In this situation,the robot is shared by all workstations,and flexible manufacturing characteristics of parts make that the residence time of the workpiece on the processing equipment changes within a given time window.Hence,the planning and scheduling of robot transportation have an important impact on the efficiency of the system.Additionally,if the automated production system is equipped with multiple handling robots on a common track,transportation resources must be cooperated to avoid the possible collisions.The actual processing time of the parts can be any value within a given time window,and the resource capacity constraints of such a system are highly coupled.Hence,the difficulty of solving the scheduling problem in automated production system with flexible processing time is significantly increased.And research progress of related issues is relatively slow,and existing research cannot meet the more complex practical application requirements.Additionally,theoretical research related to cyclic scheduling also needs to be expanded and deepened.Considering the aforementioned reasons,this research presents a comprehensive and in-depth study on cyclic scheduling problem in automated production system with fixed processing time.The main contents of this study are presented as follows.For the basic single-robot cyclic scheduling problem in automated production system,this research develops an improved genetic algorithm based on periodic sequence and infeasible solution repair strategy.This algorithm utilizes the cyclic sequence coding method and two-point crossover operation to enhance the population diversity during the evolution.In addition,an unfeasible solution repair strategy is employed where the gene fragment is adjusted adaptively within the repair process and the tabu table is used to avoid the circuitous search.Hence,this algorithm has high search efficiency and is capable to achieve high-quality solutions.To test the performance of the algorithm,the proposed algorithm is compared with the existing algorithms.The experimental results verify the effectiveness and the efficiency of the proposed algorithm.For the multi-robot cyclic scheduling problem in automatic production system,this research develops the handling operation priority method to deal with the collision conflict among robots.On the basis of the operation priority method,a mixed integer linear programming model and an intelligent algorithm are proposed to tackle this problem.The concept of priority of transport operation simplifies the construction of the robot no-collision constraints,and makes the mathematical model constructed better.Meanwhile,a swarm intelligence algorithm(imperialist competitive algorithm,ICA)is developed for this problem for the first time.Specifically,the algorithm utilizes the joint coding method: the priority sequence of handling operation and the assignment sequence of robots.On the basis of the characteristics of the problem,different crossover and mutation operations are performed on the two sequences to achieve the assimilation process of the ICA.An unfeasible solution repair strategy based on the robot assignment sequence is developed to repair the priority relationship of the handling operation.Finally,the proposed algorithm is compared with the optimization software CPLEX and genetic algorithm,and computational results on benchmark and randomly generated instances demonstrate the effectiveness of the proposed method.For multi-robot and multi-degree scheduling in automated production system with overlapped robot's running zone,this research develops a mixed integer linear programming model for multi-robot and multi-degree scheduling based on the proposed handling operation priority,which simultaneously contain the robots no-collision constraints and the workstation capacity constraints.In the modeling process,the relation between the moving sequence of multiple parts and the workstation capacity is analyzed.For the large-scale practical problem,the upper bound of the multi-degree scheduling problem is provided,and the finite overlapping method of the robot operation zone is proposed to accelerate the solution of the model.A comparative study on a number of benchmarking instances verifies the correctness and effectiveness of the model.Comparative experiments on randomly generated instances also verify the advantages of the proposed limited overlapping method.For single-robot automated mixed production system scheduling problem,this research establishes a mixed integer linear programming model with reentrant and parallel manufacturing characteristics on the basis of the transforming ideology for workstation capacity.Specifically,to handle the situation that the actual time of parts in the parallel unit may span multiple cycles,in the paper the dynamic process of parts with cycles in the parallel unit is transformed into a single static process within a cycle.And the object of machine capacity constraints is transformed from the overall machine to the independent units without changing the workstation capacity,thus simplifying the modeling process.Computational tests on the benchmark instance demonstrate that the handling operations in the parallel unit match alternatively with each other as the cycle changes,and proves the validity of the proposed methodology to formulate the machine capacity constraints.In addition,tests on a large number of randomly generated instances show that parallel manufacturing units can greatly shorten the delivery cycle of products and improve the production efficiency of manufacturing systems.The above research findings are applied to solve the scheduling problems taken from practical engineering environment.Firstly,real instance from electroplating line of plated though hole(PTH)is solved using the proposed multi-robot and multi-degree scheduling model.Afterwards,an actual instance from a full plate electroplating copper line of an enterprise is solved,where a mixed integer linear programming model for production-ratio oriented multi-robot cyclic scheduling problem is constructed.And the computational results of this instance verify the actual application value of this model.