DCMA: yet another derandomization in covariance-matrix-adaptation

In a preliminary part of this paper, we analyze the necessity of randomness in evolution strategies. We conclude to the necessity of ”continuous”-randomness, but with a much more limited use of randomness than what is commonly used in evolution strategies. We then apply these results to CMAES, a famous evolution strategy already based on the idea of derandomization, which uses random independent Gaussian mutations. We here replace these random independent Gaussian mutations by a quasi-random sample. The modification is very easy to do, the modified algorithm is computationally more efficient and its convergence is faster in terms of the number of iterates for a given precision. Categories and Subject Descriptors

• Type: Conference
• Year: 2007
• Where: GECCO
• Authors: Olivier Teytaud, Sylvain Gelly
• Keywords: Evolution Strategies, Famous Evolution Strategy, GECCO 2007, Independent Gaussian Mutations