A random search strategy for function optimization

Applied Mathematics and Computation - AMC 01/1988; 28(3):223-229. DOI: 10.1016/0096-3003(88)90138-5

ABSTRACT A random search strategy is presented which uses clustering of best function points. This method can be used to locate starting points for other algorithms and to examine the features of a function under consideration.

  • [Show abstract] [Hide abstract]
    ABSTRACT: The methodology and results of a Monte Carlo optimization of component placement on a printed circuit board are presented. The overall reliability of the component mix on the printed circuit board is the criterion of optimality. Arrhenius relations are used to quantify the effect of temperature on the component failure rate. The finite-element code ANSYS is used to determine the temperature distribution in two conductively cooled printed circuit boards. The various strategies of the Monte Carlo annealing optimization and the effect of the relevant parameters on board reliability are discussed
    Thermal Phenomena in Electronic Systems, 1990. I-THERM II., InterSociety Conference on; 06/1990