Implementation of a genetic algorithm for components optimal placement in electronic packaging

In the actual technological development context, as the electronic components become smaller and the interconnection densities continuously increases, the utilization of high-performance CAD techniques becomes not only helpful, but a necessity. An important part of the design automation is the component auto-placement, which allows the board designer to efficiently perform a difficult task without spending several hours doing it manually. In this paper the authors present an evolutional approach of the optimal components placement on the printed circuit board: a MathCad implementation of a genetic algorithm with its main stages: genetic encoding, population initialization, chromosome evaluation, selection, reproduction, mutation and convergence. Special emphasis is accorded to SMD components and high-density interconnection. Finally, the authors present experimental results and a comparative study of performances of the genetic algorithm versus standard statistical approach.