This paper exploits various meta-heuristic optimization techniques to learn PID controller parameters for nonlinear systems. The nonlinear systems considered here are well known ball and beam, inverted pendulum, and robotic arm manipulator. The gain parameters of the controllers are optimized by using two categories of meta-heuristic optimization techniques—swarm-based grasshopper optimization algorithm and particle swarm optimization and human-based, i.e., teacher learning-based optimization. Mean square error has been used to measure the performance of various algorithms. Robustness of these algorithms is studied and compared using parameter perturbation and external disturbance. There are substantial improvements in the performance of these plants using the mentioned algorithms as shown in the simulation results. A detailed comparative analysis of these algorithms has also been done.

本文利用各种元启发式优化技术来学习非线性系统的 PID 控制器参数。这里考虑的非线性系统是众所周知的球和梁、倒立摆和机械臂机械手。控制器的增益参数通过两类元启发式优化技术进行优化——基于群的蚱蜢优化算法和粒子群优化和基于人的，即基于教师学习的优化。均方误差已被用于衡量各种算法的性能。使用参数扰动和外部扰动来研究和比较这些算法的鲁棒性。如仿真结果所示，使用上述算法可以显着提高这些工厂的性能。