Abstract:Heuristic optimization has shown its superiority in handling identification problem of complicated system, for methods based on heuristic optimization do not have special requirements on model structures of the target system. In this paper, a piecewise function based gravitational search algorithm (PFGSA) is proposed and applied in parameter identification of automatic voltage regulator (AVR) system. In the proposed algorithm, a piecewise function is designed as the gravitational constant function to replace the traditional exponential equation. The piecewise function provides more rational gravitational constant to control the convergence of algorithm, and thus excellent searching ability is likely to be achieved. Moreover a new weighted objective function is proposed in the identification frame. Comparative experimental studies are conducted to test the searching ability of PFGSA and to verify the performance of proposed identification strategy, while genetic algorithm, particle swarm optimization and GSA are employed for comparison. The experimental results show that PFGSA performs the best on term of accuracy and stability in the parameter identification of AVR system, and the proposed identification strategy is effective.

Nonlinear System Identification Based on Gravitational Search Algorithm

The Identification Research of Nonlinear System Based on PSO with Fuzzy Adaptive Inertia Weight

Nonlinear System Identification Using Particle Swarm Optimization

The Hammerstein Model Identification Based on PSO

Identification of Time-Varying Delay Systems Using Particle Swarm Optimization

Nonlinear System Identification Based on Evolution Particle Swarm Optimization

Parameter Identification of Nonlinear System Model Based on Bacterial Swarm Foraging for Optimization

Identification of Nonlinear System Based on a New Hybrid Gradient-Based PSO Algorithm

Model selection and parameter estimation of nonlinear system based on PSO

Study on Method of Nonlinear System Identification

Improved genetic programming for system identification

Particle Swarm Optimization Iterative Identification Algorithm and Gradient Iterative Identification Algorithm for Wiener Systems with Colored Noise

Identification for Non-Linear Systems Based on Particle Swarm Optimization and Recurrent Neural Network

Modeling a complex nonlinear system with particle swarm optimization and parallel-optimized least squares support vector regression

Parameter Estimation of Nonlinear Systems Model Based On Hybrid Particles Swarm Optimization Algorithm

Gravitational Search Algorithm with a New Technique.

Nonlinear System Identification With Genetic Algorithms

Particle Swarm Optimal Identification of Wiener Model and a Case Study

Complex nonlinear system identification based on cellular particle swarm optimization

Stochastic Nonlinear System Identification Based on Hfc-Gp

Piecewise Function Based Gravitational Search Algorithm and Its Application on Parameter Identification of AVR System