Vector evaluated particle swarm optimization approach to solve assembly sequence planning problem
MULTI-CRITERIA Assembly sequence planning (ASP) is known as large scale, time consuming combinatorial problem. Production scheduling is a complex combined optimization problem and the optimization method of which is not perfect [1]. The product order of assembly is the main focus of ASP to deter...
| Main Authors: | , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Malaysian Society for Computed Tomography & Imaging Technology (MyCT)
2022
|
| Subjects: | |
| Online Access: | http://eprints.uthm.edu.my/7492/ http://eprints.uthm.edu.my/7492/1/J14374_b987dbf64de5e1e27c85acb5ea035ff7.pdf |
| Summary: | MULTI-CRITERIA Assembly sequence planning (ASP) is known as large scale, time consuming combinatorial
problem. Production scheduling is a complex combined optimization problem and the optimization method of which is
not perfect [1]. The product order of assembly is the main focus of ASP to determine, which is subject to precedence
constraint matrix (PM) that is to be strictly followed in the assembly line to shorten the assembly time and hence save
the assembly cost. Refs. [2, 3] proposed the concept of Assembly Precedence Relations (APRs), which is applied to
determine the precedence relations among the liaisons in the product. Cut-set analysis method by which the number of
queries can be reduced by 95% [4]. More efficient queries is proposed in ref. [5]. When number of parts increase the
problem became more complex. Heuristic methods developed to overcome this complicity. It is more efficient but it
may stick in local optima, no guarantee that global optima may be found. Some heuristic methods may use Neural
Network (NN), which need system training before start searching. Meta-heuristic method is able to escape the local
optima. Simulated Annealing (SA) is used where search is done in sequence basis and to solve optimization problems.
Ref. [6] used (SA) approach, which is based on searching via all the feasible sequences. This disadvantage is overcome
by an improved cut-set [7, 8]. Generation and evaluation of assembly plans, when the number of parts is large their
planer is slow [9]. Genetic Algorithm (GA), where the genes in chromosomes represents the components of the product
[10, 11]. An integrated approach such that liaison graph represents the physical connections between two components
[15]. An extension to previous work is proposed in [16]. Finding a method to determine global optima or near global
optima more reliably and quickly [17]. The definition of genes and evaluation criteria here are based on the connector
concept [18]. The complete or partial automation of assembly of products in smaller volumes and with more rapid
product changeover and model transition has enabled through the use of programmable and flexible automation. AI is
increasingly playing a key role in such flexible automation systems [19]. |
|---|