Frequency modulation sound parameter identification using novel hybrid evolutionary algorithms

The frequency modulation sound parameter identification is a complex multimodal optimization problem. In this paper, we proposed four evolutionary hybrid algorithms to solve this problem. First we combine genetic algorithm (GA) and queen-bee algorithm (QB) with a random optimization method (RO) and generate memetic and QB-memetic hybrid algorithms, respectively; then modified Nelder-Mead simplex algorithm (MNM) combine with particle swarm optimization (PSO) and differential evolution (DE) algorithms and generate MNM-PSO and MNM-DE hybrid algorithms, respectively. The proposed algorithms are compared in terms of three measures: success rate, average values of the cost function in all successful runs, and the minimum cost in 20 runs. The obtained results demostrate the proposed hybrid algorithms have the better performance than their non-hybrid competitors in the most times. Also The PSO has the best minimum and average value of the cost function and the MNM-PSO has the best success rate.