The Optimization Research Of Surface Mounting Machine

Electronics manufacturing industry has been developed as a symbol of nation’sscience and economy in recent years and SMT(Surface Mounted Technology) plays animportant role in it. The SMT system contains lots of equipments and the placementmachine which picks components from feeder and places them on PCB(Printed CircuitBoard)is the most important. The placement machine would take more time than otherequipments to complete a PCB, so it’s necessary to research the placement optimization.Firstly, this paper introduces SMT system as well as working process of placementmachine imply. The problem is divided into feeder assignment problem and component-sequencing problem through analysing the effect of production efficiency. A practicaloptimization model of this problem is established based on some reasonableassumptions. The placement process consists of lots of pick-and-place cycles, so thisproblem will be solved by two steps, getting pick-and-place cycles’ mounting positionand then optimizing these cycles’ sequencing which can be regarded as TSP(Travelingsalesman Problem).If the current nozzle didn’t fit the component, the placement head has to bechanged in the station after moving to it,and the nozzles should be assigned before theplacement machine working. Considering that It will take a long time to change anozzle, the changing time has great influence on efficiency of production. The choice ofnozzles should minimize the number of changing nozzles, Furthermore, the number ofpick-and-place cycle influences the placement time directly and nozzles should beassigned according to the load balance of placement heads. Placement head moves fromPCB to feeder constantly and feeders’ locations influence the moving time. This paperproposes an algorithm to solve those problems excellently. This problem is transformedto TSP successfully through getting pick-and-place cycles. Nearest neighbor algorithmis utilized to optimize the placement sequence in cycles which have lots of mountingcomponents.Through learning the traveling salesman problem-based，decided to use genetic ant colony algorithm to solve the problem. The crossover and mutation probability oftraditional genetic algorithm is fixed, designing an adaptive strategy to improve thealgorithm, and then fuse improved algorithm with ant colony algorithm. General fusionpoint is fixed, in order to achieve better integration of results, according the algorithms’speed-time graph design a dynamic fusion methods. Through matlab simulation resultsshow that the improved genetic algorithm-ant colony algorithm can achieve the desiredeffect for this problem. At last, adopting VC6.0as platform and using C++resolve thisoptimization algorithm, through the MFC displaying the optimized placement sequence.