基于遗传算法的多变量增量型广义预测控制  

Multivariable Incremental Generalized Predictive Control Based on Genetic Algorithm

作　　者：李奇安[1] 邵明新[1] 李悦[1] 

机构地区：[1]辽宁石油化工大学信息与控制工程学院,辽宁抚顺113001

出　　处：《江南大学学报（自然科学版）》2013年第5期626-630,共5页Joural of Jiangnan University (Natural Science Edition)　

摘　　要：针对多变量增量型广义预测控制算法增量系数难于根据系统特性选择的问题,提出基于遗传算法的多变量增量型广义预测控制算法。多变量增量型广义预测控制算法通过将求解出来的控制增量分别乘以相应的增量系数,来提高系统的鲁棒性和稳定性,同时采用阶梯式的控制策略降低多变量广义预测控制算法的计算量,运用改进的遗传算法优化整定多变量增量型广义预测控制的增量系数增强系统的跟踪特性。通过原油常压塔的侧线产品质量控制的仿真,验证了该方法有较好的鲁棒性和抗干扰性。A muhivariable incremental generalized predictive control using the genetic algorithm is developed to get incremental factors according the systemg performance.The control increment derived from the conventional generalized predictive control is multiplied by a factor to obtain the real increment to enhance the control systemg robustness and stability,and a stair-like generalized predictive control strategy is applied to reduce computational load.At the sametime,genetic algorithm is used to optimize the factor to improve the system tracking capability.The simulation result of a atmospheric distillation column siding quality control shows that the proposed algorithm has better robustness and anti-disturbance performance than normal multivariable generalized predictive control.

关 键 词：多变量广义预测控制 增量系数 遗传算法 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]