|Robust RST Control Design based on Multi-Objective Particle Swarm Optimization Approach
Riadh Madiouni, Soufiene Bouallègue*, Joseph Haggège, and Patrick Siarry
International Journal of Control, Automation, and Systems, vol. 14, no. 6, pp.1607-1617, 2016
Abstract : "In this paper, a novel method for the digital two-Degrees-Of-Freedom (2DOF) controller design, called
canonical RST structure, is proposed and successfully implemented based on a Multi-Objective Particle Swarm
Optimization (MOPSO) approach. This is a polynomial control structure allowing independently the regulation
and the tracking of discrete-time systems. An application to the variable speed control of an electrical DC Drive
is investigated. The RST design and tuning problem is formulated as a multi-objective optimization problem.
The proposed MOPSO algorithm which is based on the Pareto dominance is used to identify the non-dominated
solutions. This approach used the leader selection strategy that is called a geographically-based system. In addition,
the adaptive grid method is used to produce well-distributed Pareto fronts in the multi-objective formalism. The
well known NSGA-II and the proposed MOPSO algorithms are evaluated and compared with each other in terms
of several performance metrics in order to show the superiority and the effectiveness of the proposed method.
Simulation results demonstrate the advantages of the MOPSO-tuned RST control structure in terms of performance
Keyword : DC drive, digital RST controller, external archive, multi-objective design, NSGA-II, particle swarm optimization, Pareto front, robustness constraints.