Simulated annealing as a global optimization algorithm used in numerical prototyping of electronic packaging

There are many optimization algorithms, which can be used to find the extremum of a function. However, in optimizing the function, which is obtained from the numerical calculations, it is necessary to apply the proper global optimization algorithm. It is caused by the high nonlinearity of the numerical function's response. The nonlinear function with not known analytical form may have many local extrema and only one global extremum. Such problems occur in numerical prototyping, when the finite element method is used, for example in numerical optimization in electronic packaging in order to inmprove the component's reliability. There is a group of optimization methods, which can solve such nonlinear problem with success. These methods are capable to avoid of being trapped in local extemum. To this group belongs: simulated annealing, evolutionary algorithms, tabu search, ant search, etc. In this paper the simulated annealing algorithm is described and applied to optimize the electronic passive elements.