The effect of solution structure on the performance of non-derivative optimization algorithm

This paper analyses the performance of solution structure on non-derivative optimization algorithm based on characteristic of probability density function with Gaussian distribution. Using hybrid algorithm of non-derivative with two kinds of different solutions expression, the digital simulation experiments are made for benchmark functions. The results show that the structure form of solution of non-derivative optimization algorithm has an important influence on the accuracy and speed of convergence for function optimization. The function with different global minimum should adopt different structure of solutions expression. The non-derivative optimization algorithm with solutions expression of Gaussian distribution has good adaptability.