Evolutionary genetic approach to determine junction temperature in electronic packages

Genetic algorithms are adaptive search algorithms based on the Darwinian principle of natural selection and survival of the fittest. They efficiently exploit data at hand to anticipate new search points with expected improvement in performance. The aim of this paper is to show the feasibility of applying genetic algorithms to determine the thermal profile of a basic package. To illustrate this, genetic algorithms were employed to compute the junction temperature of a 12-pin lead DIP chip. Furthermore, the package parameters were modified to check the capabilities of genetic algorithms to determine the junction temperature of the chip. The paper also shows that the solutions obtained from this evolutionary search method do not depend on the initially generated members of the family of solutions. This application is specific only for purposes of illustration. This study also elaborates the factors that affect the efficiency of the solutions obtained by genetic algorithms. Finally, this paper shows that the rules of natural selection can be used to solve more complex problems faced in the packaging industry