We evaluate two Evolution Strategy-based optimisation algorithms that are known to perform well with multi-modal fitness landscapes, comparing them with each other and a standard Evolution Strategy. We apply these algorithms to a simple FM synthesiser timbre matching problem, which exhibits rugged multi-modal characteristics. All three algorithms are shown to be capable of finding globally optimal solutions. The Fuzzy Clustering Evolution Strategy is both computationally expensive and slow to converge. However, it produces globally optimal results with very high probability, compared with the other two algorithms. In contrast, whilst both the Evolution Strategy and the Cooperative Co-Evolution Strategy are significantly less computationally expensive, the Cooperative Co-Evolution Strategy is significantly quicker to converge than the other two.
Mitchell, T. J. (2006). A comparison of evolution-strategy based methods for frequency modulated musical tone timbre matching. In I. Parmee (Ed.), Adaptive Computing in Design and Manufacture: Proceedings of the Seventh International ConferenceInstitute for People-centred Computation (IP-CC)