Craziness based and Velocity Relaxed Swarm Optimized Intelligent PID Controlled AVR System

This paper explores a comparative performance study of two new classes of particle swarm optimization (PSO) techniques and binary coded genetic algorithm (GA) applied to the optimization of proportional-integral-derivative (PID) gains of PID-controlled automatic voltage regulator (AVR). The two novel swarm optimization techniques are velocity update relaxation particle swarm optimization (VURPSO) and craziness based particle swarm optimization (CRPSO). Incorporation of velocity-updating relaxation strategy in conventional PSO reduces computational effort and enhances searching ability in VURPSO. Enhanced searching ability in normal PSO is also observed in CRPSO by inclusion of a new velocity updating strategy and craziness. In comparative study, it has been revealed that VURPSO exhibits better transient performance than CRPSO. GA yields suboptimal results. For on-line, off-nominal system operating conditions Takagi Sugeno fuzzy logic (TSFL) has been successfully applied to obtain on-line responses.