The methodology and results of an optimization of placement of components on a printed circuit board using Simulated Annealing method (SA), Tabu Search (TS), and Genetic Algorithm (GA) are presented. The overall reliability of a component on PCB is the objective function. Arrehenius relations are used to quantify the effect of temperature on the component failure rate. The finite element code ABAQUS is used to determine the approximated green function of heat equation for calculating the temperature distribution in a conductively cooled printed circuit board. The effects of the relevant parameters on board are discussed. Examples are presented to show the performance of the numerical heat conduction model and to compare the performance of SA, TS, and GA.