Dynamic inertia weight artificial bee colony versus GA and PSO for optimal tuning of PID controller

Abstract

Artificial bee colony (ABC) algorithm is one of the most recently used optimisation algorithms. ABC algorithm has been extracted from the intelligent behaviour of honeybees swarm. Artificial bee colony has limitations due to its stochastic searching characteristic and complex computation that result in slow convergence to the global optimum solution. Inertia weight-based ABC method is introduced to improve the performance of the standard ABC algorithm by controlling the effect of the initial population in the new expected solution which leads improvement in the convergence rate of ABC. This work compares the performance of inertia weight ABC algorithm with the standard ABC, PSO, and GA algorithms in tuning the proportional-integral-derivative (PID) controllers for the ball and hoop system which represents a system of complex industrial processes, known to be nonlinear and time variant. Simulation results show that inertia weight ABC algorithm is highly competitive, often outperforming PSO and GA algorithms and achieving convergence rate better than the standard ABC.