A Differential Evolution Optimization Approach To Optimize The Pick-and-Placing Problem

The rapid development and popularization of Surface Mount Technology(SMT) has played an unique role in promoting the development of modern information industry, and SMT has become one of the essential technologies to manufacture modern electronic products. A surface mount-placement machine, whose assembly planning is the bottleneck of the productivity improvement of SMT system, is the most important equipment of the whole system. In this paper, optimization of SMT machines mount technology has been studied and the intention is to propose an optimization method which has a better performance, lower complexity and higher practicability, so that we can reduce the overall assembly time and improve the system productivity.In this paper, the assembly planning of machines was analyzed, an integrated optimization model was founded according to the factors we induced that affect the placement efficiency.Then the differential evolution algorithm (DE) applied to PCB-mount technology optimization, and in accordance with the related convergence of Markov chain theory to prove that under ideal circumstances its convergence with probability1convergence. Improved DE algorithm (MDE) with a migration operation, in order to increase the population diversity.For to get the PCB’s CAM data from PCB’s CAD design, This paper studies the data obtain and conversion. As the nozzle to replace need a long time and have a great impact on the placement process optimization, this paper studies the nozzle selection and assignment. Data obtain and conversion used mainly CAM350compatible with AUTOCAD approach. In order to minimize nozzle changing operation, the number of selected nozzle types was minimized. In order to use all the heads efficiently, head’ workload was assigned to be balanced in most possibility. These provided a good start for placement sequence optimization later.The genetic optimization(GA) approach, DE and MDE algorithm were employed to plan PCB assembly respectively, and optimization of their results were compared.The results show that DE and MDE required time for mounting electronic parts can be decreased obviously, and they can resolve the problem more effectively than GA.