A Branch-and-bound Combined With Petri Net For Two-sided Assembly Balancing Problem

• Main Authors: Liu, Chin-June, 劉欽榮
• Other Authors: Wei, Nai-chieh
• Format: Others
• Language: zh-TW
• Published: 2013
• Online Access: http://ndltd.ncl.edu.tw/handle/08553155671977334285

碩士 === 義守大學 === 工業管理學系碩士在職專班 === 100 === This research is to develop a solution procedure for two-sided assembly line balancing problems (TALBP). A two-sided assembly line uses both (left and right) sides of the line in parallel matter where tasks can be performed simultaneously. Reported advantages from using such a TALBP are shorter the line length and the throughput time, lower the cost of tools and fixtures, as well as reduction of material handling, etc. Many kinds of algorithms had been developed to balance two-sided assembly line in recent years such as:“An ant colony optimization algorithm for balancing two-sided assembly lines”,“A station-oriented enumerative algorithm for two-sided assembly line balancing”,“A mathematical model and a genetic algorithm for two-sided assembly line balancing”. These algorithms had applied many mathematical formulations to balance two-sided assembly which is in lack of clear and simple illustration Branch-and-bound algorithm can assign tasks clearly in balancing process by using graphic branch and bound tress (sub-trees) which disadvantage is complicated enumeration process due to inefficient reduction in branching process. In order to remove aforementioned shortcomings, Petri net’s features (incidence matrix and reachability graph) were developed and could help to delete unnecessary sub-trees and assign the tasks efficiently to increase transparency of balancing process. Petri net indeed helps branch-and-bound algorithm in decision of tasks to delete sub-trees and reduce complexity of task assignments. In the branching process, Petri net is a mathematical and graphical modeling tool which monitors the behavior of system and simulates dynamical and concurrent activities of firing processes efficiently. In addition, this support can prevent from misjudgment as much as possible. This research has the same result in balancing efficiency after comparing with other algorithms.