A novel statistically tracked particle swarm optimization method for automatic generation control

Cheshta Jain,H K Verma,L D Arya

doi:10.1007/s40565-014-0083-x

Cheshta Jain, H K Verma + Show 1 more

Open Access

https://doi.org/10.1007/s40565-014-0083-x

Copy DOI

Abstract

Particle swarm optimization (PSO) is one of the popular stochastic optimization based on swarm intelligence algorithm. This simple and promising algorithm has applications in many research fields. In PSO, each particle can adjust its ‘flying’ according to its own flying experience and its companions’ flying experience. This paper proposes a new PSO variant, called the statistically tracked PSO, which uses group statistical characteristics to update the velocity of the particle after certain iterations, thus avoiding local minima and helping particles to explore global optimum with an improved convergence. The performance of the proposed algorithm is tested on a deregulated automatic generation control problem in power systems and encouraging results are obtained.

Highlights

Particle swarm optimization (PSO) is a population based stochastic search algorithm, proposed by Kennedy and CrossCheck date: 20 November 2014Eberhart
A basic PSO algorithm is developed based on social behavior of animals like bird flocking
Basic PSO works well with the simple optimization problem, but it can be trapped on local best value for some optimization problems

Summary

Introduction

Particle swarm optimization (PSO) is a population based stochastic search algorithm, proposed by Kennedy and CrossCheck date: 20 November 2014. Z–N method is suitable for online tuning including some trial and error which is not desirable These tuning methods have some disadvantages: an excessive number of rules to set the gain, inadequate closed loop dynamic response and difficulties with nonlinear system [2]. Such problems are efficiently solved by the controller optimization using various soft computing techniques (fuzzy logic, genetic algorithm, and PSO, etc.). Reference [9] presented modified variants of PSO using different low discrepancy sequences All these variations can help to carry out fast convergence, but still a multi-dimensional problem takes a large number of runs to reach an optimum value. Numerical results of eleven benchmark functions and AGC system are presented in section 6 and section 7 concludes the paper outcome

Basic PSO

Implementation of proposed STPSO

Implementation of distributed AGC problem

Result and discussion

Conclusion