An improvement of the standard genetic algorithm fighting premature convergence in continuous optimization

This paper discusses the trade-off between accuracy, reliability and comput ing time in the binary-encoded genetic algorithm (GA) used for global optim ization over continuous variables. An experimental study is performed on a large set of analytical test functions. We show first the limitations of th e "standard GA", which mostly requires a high computing time, though exhibi ting a low success rate, due to premature convergence. We then point out th e disappointing effect of carefully choosing and tuning the "classical" GA parameters, such as the code and mutation or crossover operators. Indeed, G ray coding and double crossover helped improving on speed, but did not answ er the problem of a too homogeneous population. To fight the premature conv ergence of GA, we emphasize at last two deciding alterations made to the al gorithm: an adaptive reduction of the definition interval of each variable and the use of a scale factor in the calculation of the crossover probabili ties. The enhanced OA so achieved is discussed in detail and intensively te sted on more than 20 test functions of 1-20 variables. (C) 2000 Elsevier Sc ience Ltd. All rights reserved.