An Imperialist Competitive And Genetic Hybrid Algorithm For Assembly Sequence Planning

Product assembly is one of the important parts in the process of mechanical manufacturing, and assembly sequence is the core problem in the assembly process, which affects the cost and quality of assembly, so the assembly sequence planning plays a crucial role in the assembly process, meanwhile it has strong application value in design and manufacturing industry. Assembly sequence planning is usually regarded as a large-scale, multi-constrained combinatorial optimization problem, which aims to find the optimal assembly sequence from a large number of assembly paths quickly, and determines the assembly sequence of product parts. Assembly sequence planning in the design stage can ensure the correctness of assembly scheme, and reduce the production cost and time fundamentally. According to the characteristics of complex products assembly sequence planning, this paper proposes an imperialist competitive and genetic hybrid algorithm to realize the assembly sequence planning in order to improve the computational efficiency. The main works are as follows:(1)Putting forward the simplified assembly model for sequence planning based on interference matrix, through analyzing the relevant information in the model and regarding the geometric feasibility of sequences as constraint condition to give the method of deducing assembly direction set. We consider the alteration of assembly direction, the change of assembly tool and the change of assembly type as the constraint conditions to build the fitness function, then we propose the concept of minimum assembly costs.(2) Based on the advantages of the imperialist competitive algorithm and the genetic algorithm, a hybrid algorithm for sequence planning is proposed by the organic combination of the two algorithms. Since the imperialist competitive algorithm is not affected by the impact of the initial sequence,it serves as the initial algorithm in the hybrid algorithm. With setting corresponding number of iterations, it will terminate the algorithm to obtain the feasible sequence as the initial sequence of genetic algorithm to ensure the feasibility of the initial population of the genetic algorithm, then it can accelerate iterative convergence in the solution space. When the hybrid algorithm achieves the termination conditions, the optimum quality sequence will be output. At the same time, the basic processes and detailed steps of the hybrid algorithm to solve the optimization sequence is put forward.(3)Carrying on the sequence planning test by the proposed hybrid algorithm with a cap assembly which contains eight parts. The imperialist competitive algorithm and genetic algorithm are also used to solve the optimal assembly sequence in the same experimental environment. Through MATLAB simulation, the optimization process and the results of each algorithm are compared and analyzed, which proves that the hybrid algorithm has good performance in solving the sequence planning problem.(4)Using 3DCS tolerance analysis software to create an instance of the assembly model, through virtual assembly based on Monte Carlo technique, respectively, the different algorithms to derive the sequence corresponding assembly simulation and key assembly characteristics of geometric deviation computing. According to the final output of the statistical data and the results of the distribution of the deviation to evaluate the accuracy of the sequence of the sequence, the guiding significance of the proposed hybrid algorithm to the actual assembly engineering is verified.