MPSO-based PID parameter optimization with novel evaluation function for typical control systems

The conventional tuning cannot respond to the performance criteria of a system comprehensively and flexibly, yet the numerical optimization methodologies, such as genetic algorithm and simplex, have shortcomings on strategy parameter selection or initial condition sensitivity. This paper propose a novel comprehensive evaluation function concerning system performance, and optimize the PID controller parameters by using Modified Particle Swarm Optimization (MPSO). Two simulation show that the proposed method performs better than conventional tuning, the design of controller can satisfy various goals, and indicates well performances under either low-order or high-order circumstances, which proved to be possessing high practical value.