In this thesis, we consider operational optimization problems of gantry-type surface mounting machines with multi-heads. We decompose the problems into hierarchical related problems, component allocation, feeder arrangement, work cycle formation, pickup sequencing, and mounting sequencing problems. Since all problems in the hierarchy are combinatorial problems and computationally intractable, we developed effective algorithms based on genetic algorithm, column generation algorithm, and branch-and-price algorithm.
First, we address the component allocation and feeder arrangement problems arising in a dual-gantry multi-head surface mounting machine. The component allocation decision is to determine which components are mounted by each gantry and the feeder arrangement decision is to determine how the feeders are arranged on the feeder slots of a feeder rack. Since multi-head gantry-type machines have distinct features of simultaneous pickup, feeder interference, we need to develop unique solution approaches for the problems. We propose a genetic algorithm of optimizing the two decisions simultaneously. The two decisions are optimized by maximizing the number of simultaneously picked up components for each access of a multi-head module, or equivalently minimizing the number of pickups, and balancing the workload between the two gantries. We propose a gene encoding method that incorporates interference between the feeders of different widths. In order to evaluate the two decisions using a fitness function based on the gantry workload, we propose a greedy heuristic for the work cycle formation and pickup sequencing decisions.
Second, we consider the work cycle formation problem in multi-head gantry-type machines. The work cycle formation decision is to determine which components are picked up for each pick-and-mount work cycle. A multi-head surface mounting machine fills up multiple heads with components through one or multiple pick-up operations before starting mou...