Development of Genetic Algorithm Procedure for Sequencing Problem in Mixed-Model Assembly Lines

• Main Author: Noroziroshan, Alireza
• Format: Thesis
• Language: English; English
• Published: 2009
• Online Access: http://psasir.upm.edu.my/id/eprint/7830/1/ABS_%3D%3D%3D%3D__FK_2009_12.pdf; http://psasir.upm.edu.my/id/eprint/7830/

One of the most important issues for manufacturing systems is to determine the optimal job sequence over the production period. Mixed model assembly line is a kind of manufacturing systems which is able to deal with variable market demand. In this research, an effective utilization of mixed-model assembly line is considered as problem statement through implementing different production strategies. The problem under study contains set of mixed-model assembly line where finding the optimal job sequence based on different production strategies is the objective of this research. Different production strategies have different objectives to be met, meanwhile the sequence of jobs can be varied based on different production strategies. The main contribution of the study was implementing four production strategies in mixed-model assembly line problems, so the company can take advantage of proposed production model in different situations to meet the challenges. The first production strategy aims to minimize the make span of assembly lines and release the products to the market as soon as possible. The second production strategies attempts to minimize the make-span, and also balancing the assembly lines. It helps to balance the workload among all assembly lines. Minimizing the variation of completion time is also considered as third production strategy. The last production strategy aims to provide ideal status for assembly lines by minimizing the make-span and variation of completion time, and balancing the assembly lines. Due to NP-hard nature of sequencing problem in mixed model assembly line, a genetic algorithm is applied to cope with problem complexity and obtain a near optimal solution in a reasonable amount of time. All data is taken from literature and the result obtained from genetic algorithm procedure for the first production strategy is compared to study mentioned in literature which represents an improvement of 5% in shortening the make-span for one set of product. For the rest of production strategies, simulated annealing algorithm is applied to check the well performance of proposed genetic algorithm through reaching the same solutions for each production strategy. In all production strategies both GA and SA reaches to the same job sequence and same value of objective functions. It confirms that the proposed genetic algorithm procedure is able to tackle the problem complexity and reach to optimal solutions in different production strategies.