RBF神经网络与粒子群协同优化的DMC算法
DMC Algorithm with Collaborative Optimization of RBF Neural Network and Particle Swarm Optimization

针对DMC算法难以精确捕捉和预测复杂系统动态行为的问题，提出一种基于神经网络与粒子群双阶段协同优化DMC算法的策略。首先，在第一阶段中，设计一种基于均值聚类算法和人工鱼群算法的LM-AFSA-RBF神经网络预测模型，利用均值聚类算法确定网络基函数的中心，人工鱼群算法进一步更新网络参数，进而对高度非线性系统进行建模，预测系统的未来输出值，优化DMC算法中原预测模型部分的输出预测值，避免非线性系统下出现模型失配的问题；其次，在第二阶段中，采用引入自适应调整策略的粒子群辨识DMC算法中的控制参数，缩短参数寻优时间，提高系统的响应速度和控制精度；最后，对改进后的DMC预测控制算法进行仿真测试，实验结果验证了双阶段优化策略的可行性和有效性。
Aimed at the problem that it is difficult for the DMC algorithm to accurately capture and predict the dynamic behavior of complex systems, a strategy based on the two-stage collaborative optimization of the DMC algorithm with neural networks and particle swarms is proposed. Firstly, in the first stage, a LM-AFSA-RBF neural network prediction model based on mean clustering algorithm and artificial fish swarm algorithm is designed, the mean clustering algorithm is used to determine the center of the network basis function, and the artificial fish swarm algorithm further updates the network parameters, the network is then used to model highly nonlinear systems, predict the future output values of the system and optimize the output predictions of the original predictive model part of the DMC algorithm, to avoid model mismatch under nonlinear systems; secondly, in the second stage, the particle swarm introducing adaptive adjustment strategy is used to identify the control parameters in the DMC algorithm to shorten the parameter optimization time and improve the response speed and control accuracy of the system; finally, simulation and test are carried out on the optimized DMC predictive control algorithm, the experimental results verify the feasibility and effectiveness of the two-stage optimization strategy.