For some decades, the assembly of printed circuit boards (PCB), had been thought to be an ordinary example of mass production systems. However, technological factors and competitive pressures have currently forced PCB manufacturers to deal with a very high mix, low volume production environment. In such an environment, setup changes happen very often, accounting for a large part of the production time.
PCB assembly machines have a fixed number of component feeders which supply the components to be mounted. They can usually hold all the components for a specific board type in their feeder carrier but not for all board types in the production sequence. Therefore, the differences between boards in the sequence determines the number of component feeders which have to be replaced when changing board types. Consequently, for each PCB assembly line, production control of this process deals with two dominant problems: the determination for each manufacturing line of a mix resulting in larger similarity of boards and of a board sequence resulting in setup reduction. This has long been a difficult problem since as the number of boards and lines increase, the number of potential solutions increases exponentially.
This research develops an approach for applying Genetic Algorithms (GA) to this problem. A mathematical model and a solution algorithm were developed for effectively determining the near-best set of printed circuit boards to be assigned to surface mount lines. The problem was formulated as a Linear Integer Programming model attempting to setup reduction and increase of machine utilization while considering manufacturing constraints. Three GA based heuristics were developed in order to search for a near optimal solution for the model. The effects of several crucial factors of GA on the performance of each heuristic for the problem were explored. The algorithm was then tested on two different problem structures, one with a known optimal solution and one with a real problem encountered in the industry. The results obtained show that the algorithm could be used by the industry to reduce setups and increase machine utilization in PCB assembly lines. / Graduation date: 1998
Identifer | oai:union.ndltd.org:ORGSU/oai:ir.library.oregonstate.edu:1957/33935 |
Date | 03 December 1997 |
Creators | Capps, Carlos H. |
Contributors | Paul, Brian K., Beaumariage, Terrence |
Source Sets | Oregon State University |
Language | en_US |
Detected Language | English |
Type | Thesis/Dissertation |
Page generated in 0.0017 seconds