Production planning and scheduling for printed circuit, board assembly has so far defied standard operations research approaches due to the size and complexity of the underlying problems, resulting in unexploited automation flexibility. In this thesis, the increasingly popular collect-and-place machine configuration is studied and the assembly sequence planning problem modeled as a generalized vehicle routing problem.;The proposed novel methodology constitutes of a constructive heuristic and an efficient interchange heuristic that extracts information on potentially beneficial interchanges based on a sophisticated, yet computationally tractable integer linear program. The initial solution is generated using a multi-stage heuristic algorithm that models and solves important subproblems as well-known optimization problems. The refinement heuristic is guaranteed never to let the initial solution, which is bounded in the worst case to be no more than a certain percent from the optimal, deteriorate in quality and provides an explicit trade-off between solution quality and computational time.;We provide theoretical results on the worst case analysis of the construction phase in the generalized vehicle routing context under the full-load, partial-load, and the partial-load with waiting cost scenarios.;Extensive computational results are also presented demonstrating how board assemblers can use the developed technique to optimize the process for their set-ups and to establish the feasibility as well as effectiveness of the overall heuristic and comparing various strategies used in the design of the heuristic. We also rigorously study the sensitivity of the assembly time on such machines with respect to the hoard parameters and machine configuration allowing equipment manufacturers to focus on improving and customizing machine design.
展开▼
