利用基因演算法解決印刷電路板人工插件之之生產線平衡問題

Abstract

在印刷電路板組裝的製程中，部份元件的組裝已由自動化設備取代人力，但仍然有相當數量且種類繁多的雙列直插式封裝(Dual In-line Package, DIP)元件需藉由人工的方法組裝於印刷電路板上。DIP的組裝生產線由多個工作站所組成，每個工作站負責置放數個元件，而由於以人工方式置放元件，影響生產線的效率與組裝品質的因素，除了有傳統的生產線平衡問題(Assembly Line Balancing Problem, ALBP)外，還需考量人員在組裝過程的順暢度，以及為避免組裝錯誤而加入的組裝限制。本研究以DIP組裝線為研究對象，設計一套基因演算法(Genetic Algorithm, GA) ，在給定工作站數量與元件的指派限制之下，將所有元件指派於工作站中，使各工作站中的最大週期時間最小，並符合各項指派限制。本研究利用台灣某電腦主機板製造廠提供的產品資訊測試本研究所設計的演算法，結果發現本研究所設計的方法所產生元件指派方式，與該公司現行之指派方式相比，能有更短的最大週期時間，並更能符合該公司所設計的指派限制。本研究亦另行模擬了更複雜的產品資訊，在以本研究所提出的方法指派元件後，亦能順利地於短時間內找出能夠符合指派限制，並使最大週期時間最小的元件指派方式。

In the assembly line for printed circuit board assembly (PCBA), automation is a tool that is being used instead of using actual labor. However, while assembling Dual In-line Packages onto PCBs, actual labor is still required. While there are many tasks on the assembly line, assigning components is what affects the production efficiency on the assembly line. This is an Assembly Line Balancing Problem (ALBP). This research is conducted regarding assembling DIP’s onto PCB’s based on assigning the components onto a fixed number of assembly stations, in order to minimize the cycle time. This is a Type II Problem where this research applies Genetic Algorithm (GA) for solving. The GA contains constraints to be solved within a short time frame with respect to these constraints. This research uses both product information obtained from an anonymous motherboard manufacturer in Taiwan and also simulated product information in the GA. The outcome of this research met the constraints and minimized the cycle time.