Research On Multi-manned Collaborative U-shaped Assembly Line Balancing Problem Based On Enhanced Discrete Gray-wolf Cuckoo Search Algorithm

With the development of China’s industrialization process and the improvement of scientific and technological level,the types of products produced by manufacturing industry are more diversified.Compared with simple products that can be produced in assembly line,the allocation of assembly tasks of small batch and diversified large-scale equipment is more difficult.How to reasonably allocate the assembly tasks of each part to each workstation has become an important factor of this kind of products to reduce production costs and improve production efficiency.The assembly process allocation of this kind of equipment is generally complex,which is reflected in many aspects.On the one hand,some complex processes need to be processed by multiple operators,on the other hand,different processes need different kinds of operators to process.Therefore,in the selection of assembly line for this product,multimanned assembly line is generally selected.At present,the linear workstation layout is mainly adopted in the research on multi-manned assembly line balancing problem.This layout has poor production flexibility,low production efficiency and large floor area.In order to improve this feature,this paper combines the multi-manned assembly line with the U-shaped Assembly Line(UAL)with bidirectional process allocation ability and strong flexibility.At the same time,considering the above-mentioned multi worker collaborative processing and the needs of including multiple types of work,this paper proposes and studies the Multi-manned Collaborative U-shaped Assembly Line Balancing Problem(MCUALBP)with kinds of worker.The main contents of the study include:(1)The characteristic analysis and modeling of the MCUALBP are researched.The number of workstations,the number of workers,and the workload balance of workers are combined to design the objective function.Different from the general multi-manned assembly line,the problems studied in this paper have the following characteristics: first,the number of types of worker has changed from one to many,and workers can only process the processes of the corresponding types of himself.How to improve the working efficiency of different types of worker has become a difficulty.Second,some processes require multiple workers to form working groups for collaborative processing,and we need to consider how to assign workers to appropriate working groups on the premise of ensuring work efficiency Third,the U-shaped workstation layout increases the difficulty of two-way allocation of processes to the process scheduling.Combined with the above characteristics,this paper comprehensively considers the distribution of workers and processes,and establishes a set of mathematical model for the multimanned collaborative U-shaped assembly line with kinds of workers.(2)The cuckoo algorithm is discretized and improved,and an enhanced discrete cuckoo algorithm is proposed(EDCS).Because the proposed problem belongs to NP hard problem,there will be a combination explosion when using accurate algorithm to solve large-scale problems,and it is difficult to obtain a better solution in a reasonable time,so this paper chooses meta heuristic algorithm to solve it.Meta heuristic algorithms are sometimes prone to fall into local optimization,and solving complex problems in the search domain requires the algorithm to have the ability to jump out of local optimization.As a bionic algorithm,cuckoo search algorithm has the characteristics of strong global search ability and simple parameters,which can be used as an algorithm to solve the problem proposed in this paper.In this paper,the cuckoo algorithm is discretized and enhanced.This paper mainly improves the cuckoo algorithm from two aspects.First,it improves the generation mode of the initial population,describes the importance of the process in the assembly process by introducing the concept of bottleneck degree,and affects the priority of process allocation in the generation process of the initial individual;Second,in view of the lack of neighborhood search strategy for elite individuals,this paper performs local search on elite individuals under the constraints of process order,and implements a neighborhood 2-opt search operator for elite individuals,which improves the local search ability of the algorithm for elite individuals.In addition,this paper designs a decoding scheme based on the minimum free time of workers,giving priority to the idle workers who are close to the starting time of the current allocation process.(3)Combining the enhanced discrete cuckoo algorithm with gray-wolf algorithm,an Enhanced Discrete Gray-wolf Cuckoo Search algorithm is proposed(EDGCS).The search direction of cuckoo algorithm only points to the optimal individual of the current population,means the search direction is single,which is not conducive to a comprehensive search of individual neighborhoods.When the gray wolf algorithm recombines individuals,it exchanges information with multiple better individuals,which has the characteristics of diversified search directions.Combining it with the nesting operator based on Levy flight in the cuckoo algorithm,a Gray-wolf-Levy nesting operator is proposed,so that the recombination direction of individuals is changed from a single optimal body to multiple better individuals,so that the optimization direction of the cuckoo algorithm is variable,It is beneficial to enhance the global search ability of the algorithm.(4)Numerical experiments are carried out on the improvement of cuckoo algorithm to prove the effectiveness of the improvement.In order to prove that EDCS can improve the optimization efficiency of CS,for the test cases of the standard U-shaped assembly line balancing problem,on the basis of CS,the initial solution generation strategy and the elite neighborhood search operator are introduced respectively,and comparative experiments are carried out to verify the accuracy and stability of EDCS.In order to prove the effectiveness and superiority of EDGCS,a comparative experiment is carried out with EDCS on the test case of the standard U-shaped assembly line balancing problem,which proves that EDGCS is more effective.A comparative experiment is carried out with the algorithm proposed in other literatures on the test case of the standard U-shaped assembly line balancing problem,The superiority of EDGCS is verified.At last,aiming at MCUALBP studied in this paper,EDGCS is compared with EDCS and the enhanced genetic algorithm(ECS),which verifies the effectiveness and superiority of EDGCS in the problem studied in this paper is better.