Designing of Controller: a Multi-objective Genetic Optimization Approach
Md. Ahsan Habib, Mohammad Manjurul Islam, M. S. Alam · Dhaka University Journal of Science · 2011
Controller design for any system requires multiple objectives to be minimized or optimized. This involves the satisfaction of multiple objectives which are conflicting such as Rise time and Maximum overshoot. It is difficult to meet these two conflicting objective simultaneously using conventional design methods. This paper presents an investigation on design and development of Multi-Objective Genetic algorithm (MOGA) based controller to achieve that goal. Two aerial vehicle systems have been investigated in this work. Performance of conventional technique such as Ziegler-Nichols (Z-N) method has been compared and analyzed with the intelligent tuning technique MOGA. Simulation results confirm the superiority of this MOGA approach over conventional techniques. Moreover, it is proposed that if an objective value does not satisfy its corresponding goal value, a penalty can be imposed onto its competing objective in proportion to the extent of violation. This approach enhance the generation of the desired goal solutions.