Nature inspired optimization algorithms have made substantial step towards solving of various engineering and scientific real-life problems.Success achieved for those evolution-ary optimization techniques are due to s...Nature inspired optimization algorithms have made substantial step towards solving of various engineering and scientific real-life problems.Success achieved for those evolution-ary optimization techniques are due to simplicity and flexibility of algorithm structures.In this paper,optimal set of filter coefficients are searched by the evolutionary optimiza-tion technique called Opposition-based Differential Evolution(ODE)for solving infinite impulse response(IIR)system identification problem.Opposition-based numbering con-cept is embedded into the primary foundation of Differential Evolution(DE)technique metaphorically to enhance the convergence speed and the performance for finding the optimal solution.The population is generated with the evaluation of a solution and its opposite solution by fitness function for choosing potent solutions for each iteration cycle.With this competent population,faster convergence speed and better solution quality are achieved.Detailed and balanced search in multidimensional problem space is accomplished with judiciously chosen control parameters for mutation,crossover and selection adopted in the basic DE technique.When tested against standard benchmark examples,for same order and reduced order models,the simulation results establish the ODE as a competent candidate to others in terms of accuracy and convergence speed.展开更多
文摘Nature inspired optimization algorithms have made substantial step towards solving of various engineering and scientific real-life problems.Success achieved for those evolution-ary optimization techniques are due to simplicity and flexibility of algorithm structures.In this paper,optimal set of filter coefficients are searched by the evolutionary optimiza-tion technique called Opposition-based Differential Evolution(ODE)for solving infinite impulse response(IIR)system identification problem.Opposition-based numbering con-cept is embedded into the primary foundation of Differential Evolution(DE)technique metaphorically to enhance the convergence speed and the performance for finding the optimal solution.The population is generated with the evaluation of a solution and its opposite solution by fitness function for choosing potent solutions for each iteration cycle.With this competent population,faster convergence speed and better solution quality are achieved.Detailed and balanced search in multidimensional problem space is accomplished with judiciously chosen control parameters for mutation,crossover and selection adopted in the basic DE technique.When tested against standard benchmark examples,for same order and reduced order models,the simulation results establish the ODE as a competent candidate to others in terms of accuracy and convergence speed.
