Application of genetic algorithms for electronic devices placement in structures with heat conduction through the substrate

The paper presents an implementation of a genetic algorithm for electronic devices placement optimization. It is assumed that nine electronic devices can be positioned in any place on a substrate surface (in accordance to a grid). Because of that, a new, 2-level optimization algorithm has been created. Positions of nine electronic devices each dissipating different power are optimized in respect of thermal criteria and wiring length. This paper compares results obtained using simple algorithm having permutational coding with new 2-level algorithm.