An Particle Swarm Optimization Approach for Assembly Sequence Planning

Abstract

Assembly sequence planning (ASP) is the foundation of the assembly planning which plays a key role in the whole product life cycle. Although the ASP problem has been tackled via a variety of optimization techniques, the particle swarm optimization (PSO) algorithm is scarcely used. This paper presents a PSO algorithm to solve ASP problem. Unlike generic versions of particle swarm optimization, the algorithm redefines the particle's position and velocity, and operation of updating particle positions. In order to overcome the problem of premature convergence, a new study mechanism is adopted. The geometrical constraints, assembly stability and the changing times of assembly directions are used as the criteria for the fitness function. To validate the performance of the proposed algorithm, a 29-component product is tested by this algorithm. The experimental results indicate that the algorithm proposed in this paper is effective for the ASP.

Multi-objective discrete particle swarm optimisation algorithm for integrated assembly sequence planning and assembly line balancing

“…Various versions of PSO algorithm have been proposed to optimise multi-objective problems for independent ASP and ALB. 28,30,33,39,40 One of the PSO versions for optimising multi-objective problems is known as DPSO that was first proposed by Rameshkumar et al 31 for scheduling problem and later adopted to optimise ASP problem. 40 However, in these works, the multi-objective problem was handled by bundling all objectives into a single objective that leads to only one solution.…”

A review on assembly sequence planning and assembly line balancing optimisation using soft computing approaches

“…Wang and Liu [22] categorised the stability strengthened by the assembly connectors into three levels according to connection strength; strong, weak and unstable con nection. This classification is also used by [25]. In this approach, strong connection will be assigned '0' index, weak connection '1' index and unstable connection '2' index.…”

“…Therefore, this attribute is not included as objective. However, in a small number of ASP research as in [25,[29][30][31], the geometrical constraint is declared as one of ASP objective. When the assembly geometry is unfeasible, the large penalty index will be given to fitness function.…”

“…Not many papers that applied PSO to ASP and ALB have been published. This has motivated researchers [8,25,82] to apply PSO to ASP and ALB optimisation. Besides that, PSO is a simple algorithm because it only uses a single velocity formula to evolve [60].…”

“…Another important issue of original PSO is that it is easily trapped in local optima [22]. To solve this problem, [25] introduced a new mechanism of updating velocity by using one of two formulas randomly instead of single formula. Meanwhile, [22] introduced a chaotic operator to diversify the updated particle posit ion, which finally help to reduce premature convergence.…”