Evolutionary computation based optimization in fuzzy automatic generation control

This paper presents a comparative optimization performance and transient performance studies among three evolutionary computational techniques as genetic algorithm (GA), hybrid particle swarm with constriction factor approach (HPSOCFA) and hybrid Taguchi particle swarm optimization (HTPSO) methods in automatic generation control. For comparative study of performances, the above-mentioned techniques are firstly used to determine nominal optimal proportional-integral-derivative gains of PID controller and nominal transient responses of frequency deviations in each area for nominal area input parameters. HTPSO based optimal gains result in true optimal transient responses of frequency deviations and tie line power flow changes. HPSOCFA and specially GA yield sub optimal responses. Finally, for on-line, off-nominal input parameters, Sugeno fuzzy logic (SFL) manipulates the GA/HPSOCFA/HTPSO based off-line nominal optimal gains very fast, resulting in on-line off-nominal optimal gains and off-nominal optimal transient responses under on-line varying system parameters and load conditions