Near-Optimal Tuning of Linear Controllers Based on Genetic Algorithm and Swarm Intelligence: A Flight Control Example

This paper presents a method for the tuning o] PID-type controllers for systems with time-varying dyna.mics based on optimization by genetic algorithms (CA) and swarm intelligence, i. e. a modified version of particle swarm optimization, The powerful abilities of bio-inspired computing in locating the optimal (or nearoptimal) solutions to a given optimization problem are utilized for determinzng the parameters of the controller in order to meet specified performance objectives for a given problem in terms of weighted integral of d ifferent signals. The presented study recommends a. new modification ui PSG algorithm where a novel approach for tuning of the lcarning [actors is employed to improve the speed of convergence and the accuracy of the search in the multi-dimensional space. The proposed technique has the flexibility to be applicable to a wide range of control problems. In order to illustrate the performance of the technique. development and application of this algorithm for an aerospace launch vehicle during atmospheric flight in an experimental setting is presented.