摘要
在实际应用中,对于带参数的最优控制问题,有时当参数发生微小的扰动时,目标泛函相差甚远。针对这类敏感性问题,采用在性能指标中引入罚函数的方法,建立一类考虑参数敏感因素的最优控制问题模型,同时提出了一种新的数值算法,以避免新模型求解中涉及目标泛函二次变分计算的难题。数值仿真结果证明了模型和算法的正确性和有效性。
With regard to optimal control with parameters, sometimes the cost functional becomes quite different when there exists a minute disturbance of the parameter in practical applications. Aiming at a class of such sensitivity problems, penalty functional was introduced into the performance criteria and the optimal control model was setup considering the parameter sensitivity. At the same time, a novel numerical algorithm was proposed in order to avoid dealing with solving cost functional second-order variational calculation in the new model. Numerical simulation results prove the model and algorithm are correct and effective.
出处
《系统仿真学报》
CAS
CSCD
北大核心
2009年第11期3206-3208,3233,共4页
Journal of System Simulation
基金
教育部科技重点基金(207104)
贵州省国际合作项目(2006)400102
关键词
最优控制
敏感性
数值仿真
参数选择
梯度
optimal control
sensitivity
numerical simulation
parameter selection
gradient