Implementation of a genetic algorithm for components optimal placement in electronic packaging
Paul Mugur Svasta, H. Carstea, Marius Rangu, A. Avram · 2003
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.