LAUSR

Artificial Intelligent Optimization (AIO) algorithms learn from the past searches via using a group of individuals or agents. These Artificial Intelligence-based optimizing techniques are able to solve complex optimization problems with complicated constraints. They find the optimal in the low possible number of iterations, where optimal means the best from all possibilities selected from a special point of view. This paper presents a research on employing AIO methods with aim to Infinite Impulse Response (IIR) system modeling for design and optimization of IIR digital filters. The proposed methods cover a variety of AIO methods; algorithm based on evolution strategy (genetic algorithm) and heuristic algorithms (particle swarm optimization, population-based; gravitational search algorithm, and inclined planes system optimization, both population-based and based on Newton’s laws). In this paper, the IIR system modeling is solved as a constrained single-objective optimization problem in the Mean Squared Error (MSE) fitness function and is evaluated for two different benchmark IIR plants with high and low orders. To evaluate performance, efficiency and efficacy of the methods, two important criteria are used: “Indicator of Success (IoS)” and “Degree of Reliability (DoR)”. In addition, the effect of decreasing population size (search agents) is analyzed on the performance and efficiency of the algorithms. Simulation results clarify the success of the research in terms of the MSE, IoS and DoR.