Genetic Planner to Support DFA Assemblability Evaluation: Insights on Integrated Design and Planning for Assembly

Abstract

Abstract Several Design for Assembly (DFA) methodologies have already been introduced both in research and industry, among which methodologies having quantitative assemblability evaluations are more favored in industry. The evaluation criteria involved in such quantitative measures are only concerned with the difficulty of performing assembly operations one by one. These measures are then combined by a formula to find the total product evaluation. Accordingly, the designer can improve the total assemblability by redesigning the parts that caused bad measures. It can be observed that a major effect of cascading several assembly operations in a certain order on the remaining operations to complete the product assembly is neglected by most DFA evaluation methodologies. Actually, this effect might result in serious difficulties while generating the assembly sequence usually done at a later stage in most industrial firms. To overcome these difficulties, this paper introduces some elementary insights to the integration of assembly design and planning. A systematic feature-based method is introduced to find the precedence assembly constraints and the global evaluation constraints resulting from operations cascading orders. A modified genetic algorithm that finds the best assembly sequence based on these constraints and on the traditional DFA evaluations is then introduced to be included in the DFA procedure. This genetic planner generates a more representative evaluation of the product assemblability by considering the effects of operations orders. As a result, feedback reflecting the actual assembly process difficulties can be given to the designer to improve the product design, while simultaneously generating the best sequence to assemble the product.