Component sequencing and feeder arrangement for PCB assembly machines: Integration, models, and solutions

The extensive applicability of printed circuit boards (PCBs) in numerous contemporary electronic products has placed an unprecedented demand for PCBs. Besides, customers' changing needs such as smaller in product size and greater in function and reliability force the surface mount technology (SMT) to replace the plated-through-hole technology. Among several assembly operations in an SMT assembly line, the component placement operation is a bottleneck of the line. So, this thesis mainly focuses on the optimization of the component placement operation such that the highest productivity can be achieved.; The optimization of the component placement operation entails two inter-related problems: the component sequencing problem (i.e., which component places first, second, and so on) and the feeder arrangement problem (i.e., which feeder holds which type of components). Because of their inseparable relationship, the two problems should be integrated instead of separated, as most of the researchers did. In this project, various types of mathematical models are formulated for the integrated problems of both the component sequencing and the feeder arrangement sub-problems for the two commonly used SMT placement machines: the pick-and-place machine and the chip shooter machine. The objective is to optimize the machines' performance. The models are also verified with commercial packages. Moreover, since it is very time-consuming to solve the mathematical models for optimal solutions, a heuristic approach is a prerequisite. Herein, a genetic algorithm hybridized with several improved heuristics is designed to solve the integrated problems. It is proved to be effective and more efficient.; Apart from the optimization of the component placement operation, this thesis also studies the line assignment problem in order to assign board types to proper assembly lines such that the minimal production cost can be obtained, in conjunction with the component allocation problem in order to assign component types to right placement machines in a line so that the minimal cycle time can be spent. For the line assignment and the component allocation problems, not only comprehensive and verified mathematical models are built but also genetic algorithms are proposed. Finally, a prototype of the printed circuit board assembly planning system is developed, in which the line assignment problem, the component allocation problem, and the integrated problems for both types of placement machines are combined together, to aid process engineers to determine the decision problems arising in PCB assembly.