Research On Coevolutionary-Based Hybrid Intelligent Optimization Algorithms In SMT Production

In the production of electronic products, it is an important approach to improving production efficiency and resource utilization by using advanced optimization technology to the solve the scheduling and control problems on PCB assembly line and the equipment in the line. Therefore, it is of great theoretical significance and application value that explore and study on the related theoretical issues and optimization methods about the optimization of operations of a SMT line and its key equipment, surface mounting machine.This paper studies the operation optimization problem of surface mounting machine to minimize the assembly production time of a PCB. The optimization problem involves serval highly interrelated sub-problems, which makes it very complicated and hard to solve. Therefore, how to design effective solving method after the mathematical model for the problem is formulated to shorten the assembly time and improve production efficiency of SMT assembly line, is our goal.The paper have a systematic analysis, classification, comparison and summary about the solving ideas and methods of past research work about the optimization of SMT assembly, and analyze the operation optimization problem with the objective of minimizing the PCB assembly time for a single surface mount machine. At the same time, a solving approach based on cooperative evolutionary algorithm is proposed for this problem. We have decomposed the optimization problem into multi-sub-problems and modeled them into two stages: an integer linear programming(ILP) for the nozzle assignment problem is formulated and solved by a constructive heuristic in the first stages. A mixed integer linear programming(MILP) for feeder assignment problem and pick-and-place sequencing problem is formulated and solved by combining the multi-agent evolutionary algorithm and local search algorithms, with the idea of co-evolutionary algorithm. By using the above approach, an near-optimal solution can be obtained for the problem in short time.