An adjustable grouping genetic algorithm for the design of cellular manufacturing system integrating structural and operational parameters

• Published in: Journal of manufacturing systems Vol. 44; pp. 115 - 142
• Main Authors: Jawahar, N., Subhaa, R.
• Format: Journal Article
• Language: English
• Published: Elsevier Ltd 01.07.2017
• Subjects: Adaptive parameters; Cell formation; Cellular manufacturing system; Genetic algorithm; Grouping genetic algorithm; Cell formation; Grouping genetic algorithm; Cellular manufacturing system; Genetic algorithm; Adaptive parameters
• ISSN: 0278-6125; 1878-6642
• DOI: 10.1016/j.jmsy.2017.04.017

[Display omitted] •Proposes Cellular Manufacturing Systems (CMS) model that evolves integrated structural and operational design decisions.•Presents a non-linear & linear mathematical formulation for the proposed CMS model.•Proposes Adjustable Grouping Genetic Algorithm (AGGA) with features to adjust coding for machine duplication environment.•AGGA regulates genetic parameters towards convergence in a computationally efficient manner.•Discusses the means of extending the model and AGGA to other clustering applications. This paper presents non-linear and linear formulations for the design of a Cellular Manufacturing Systems (CMS) modeled integrating structural and operational decision parameters, and a Genetic Algorithm (GA) based on self-regulating adaptive operators. The proposed CMS model evolves the structural design decisions of number of cells, and parts – machines assignment to cells, along with operational decisions of scheduling under machine duplications and alternate routings/cross-flow environments. The distinctive features of the CMS model under consideration are: i) integration of cost elements addressing both structural and operational issues in the design of CMS; ii) capable of evolving better CMS design decisions in terms of operational cost when compared to the literature part-machine grouping decisions; iii) suitable for variety of manufacturing system designs by relaxing the model constraints. Besides, this paper proposes a new variant of Grouping Genetic Algorithm namely Adjustable Grouping Genetic Algorithm (AGGA) that has features to adjust the coding suitable for machine duplication environment of the proposed CMS model and regulate genetic parameters towards convergence. It is shown, through comparisons with Simulated Annealing (SA) algorithm, Simple Genetic Algorithm (SGA) and also optimal solutions obtained via mathematical model relaxed to fixed number of cells, that AGGA is capable of evolving optimal or near optimal solutions in a computationally efficient manner.