This paper proposes a novel Artificial Intelligence technique known as Ant
Colony Optimization (ACO) for optimal tuning of PID controllers for load frequency
control. The design algorithm is applied to a hydrothermal power system consisting of
two control areas one hydro and the other is thermal with reheat stage. To make the
system in realistic form, the system nonlinearities represented by Generation Rate
Constraint (GRC), Dead Band, wide range of parameters are introduced. Three different
cost functions have been suggested for tuning the PID controllers. The system has been
tested for various load changes to reveal the effectiveness and robustness of the
proposed technique.