We investigate the feeder assignment problem for gantry-type surface mounting devices with multiple component pickup heads and feeder interference. As the sizes of components and printed circuit boards significantly reduce, the times for picking up components from the feeders are most critical, which are greatly affected by arrangement of the feeders. Instead of time-consuming direct optimization of the cycle time, we propose a strategy of maximizing the degree of simultaneous pickups, which greatly contributes to minimizing the total cycle time. For a given feeder arrangement, we develop an efficient greedy heuristic to generate a pickup sequence that optimize the degree of simultaneous pickups. An evaluation function is developed to measure the degree of simultaneous pickups for such determined pickup sequence for a given feeder arrangement. Using the evaluation function, we then develop a genetic algorithm for searching an optimal feeder arrangement that maximizes the degree of simultaneous pickups. We also develop an extended physical-logical conversion method to apply genetic operators easily. Finally, we examine the experimental performance of the proposed algorithm.
