EVOLUTIONARY OPTIMIZATION ALGORITHM BY ENTROPIC SAMPLING

A combinatorial optimization algorithm, genetic-entropic algorithm, is proposed. This optimization algorithm is based on the genetic algorit hms and the natural selection via entropic sampling. With the entropic sampling, this algorithm helps to escape local optima in the complex optimization problems. To test the performance of the algorithm, we ad opt the NK model (N is the number of bits in the string and K is the d egree of epistasis) and compare the performances of the proposed algor ithm with those of the canonical genetic. algorithm. It is found that the higher the K value, the better this algorithm can escape local opt ima and search near global optimum. The characteristics of this algori thm in terms of the power spectrum analysis together with the differen ce between two algorithms are discussed.