Two-Stage Optimization Problem Of Multiple PCB Types Assembly

PCB is an important part of electronic products.With an increasing popularity of digital electronic products in recent years,PCB has experienced an unprecedented fast development.PCB assembly is to assemble and connect each basic electronic unit one by one on the bare board,which has become a key process in electronic manufacturing industry.Nowadays,the environment of electronic industry is high-mix and low-volume.Many production enterprises are paying more attention to how to do better in production logistics,which has become one of the most important core competitiveness in the whole industry.Surface mounting technology is applied in PCB assembly.Placement machine is the core and bottleneck equipment in the surface mounting technology.This paper takes nozzles setup and components setup into consideration,and set up two-stage optimization model problem for lowest assembly production time(including cycle time and setup time)on the assumption that the sequence of PCB assembly has been fixed.In the first stage,unit cycle time and unit setup time will be set as a constant to find the solution on nozzle loading and switching.In the second stage,cycle optimization in one turn for single PCB will be solved through setting component group and placement sequence.Finally we will get the lower total assembly production time.Since it is a NP-Hard problem,the paper designs a two-stage hybrid algorithm to optimize the total assembly production time of PCB.In the first stage,PCB grouping problem is solved by genetic algorithm.Then,the head allocation problem is solved by greedy increment method to minimize the cycle number in the same group.What's more,the nozzle setup problem between different groups is handled by KTNS algorithm for less nozzle setup.In the second stage,we apply clustering algorithm and saving mileage method to optimize the unit cycle time for single PCB by setting component group and sequence in a single turn based on the nozzle assignment and setup solution.Finally,numerical experiments are carried out through Visual Basic to examine the performance of the proposed algorithm and find that we can get a satisfied solution.All these can provide a data basis and theoretical support for the application in the future.