Research On Stochastic Mixed Model Assembly Line Balancing Problem Based On Hybrid Particle Swarm Optimization Algorithms

The assembly line balancing problem (ALBP) is an important issue in the manufacturing field; minor improvements can lead to significant gains in the economy. In the era of rapid change, the society economic activity is complicated changeful; the market competition is gradually fierce. The mixed model assembly line has got broad application in making enterprise, because it can meet customers’diversify and individuation needs, and also it has strengthened the enterprise adaptability to the marketplace. In the actual production process, operating time on the assembly line is with a lot of randomness. But in the current study, a variety of assembly line balancing design based on seting assembly work time as constant, which does not fully conform to actual production. In this paper, author take stochastic mixed-model assembly line as object of study, which have a certain amount of theoretical and practical significance. The thesis is organized as follows.1) The thesis studys the theoretical knowledge of ALBP systematically, analyzes the computational complexity of ALBP and the faxtors which can affect the balance, then gives the improvements how to improve the balance efficiency. The stochastic mixed-model assembly line balancing problem (SMMALBP) is proposed based on considering the stochastic factors which can influce the assembly operation time, and a corresponding mathematical model is established according to its characteristics.2) In this paper, an improved particle swarm optimization for ALBP of type-1according to is proposed to solve particle swrm optimization(PSO)’s premature phenomenon. The improved algorithm enhances the diversity of particles and improves the search performance by setting the priority weight. The paper also developes procedures by using Matlab, numerous calculations of rhe test problems to verify the effectiveness of the improved algorithm.3) A hybrid particle swarm optimization algorithms base on simulated annealing (SA) mechanism for solving SMMALBP. The hybrid algorithms can make the particles escape from local optima in time and achieve parallel search which can expand the search space, enhance particles’exploration and development capabilities in the solution space. So the hybrid algotithms improve the global search capability, optimize efficiency and rubustness. A large number of examples demonstrate the effectiveness of the hybrid algorithms.4) Bacause of the randomness of assembly task’s working time can lead to the workstation time become a random variable, the operator can complete all the assembly tasks in the given cycle time with a certain probability. So in this paper, a method to change the probability of completing all the workstation tasks by changing the default overrun paobability is proposed. Many examples show the feasibilityand effectiveness of this method to solve SMMALBP.5) Using the proposed PSO-SA algorithm to improve a company’s engine assembly line, which improve the original assembly line efficiency significantly and make the workstation load more balanced. The results show that the improvements are obvious.