Analysis and Synthesis of Control Systems via Risk Based Optimization

从理论和应用两方面研究了基于风险优化的控制系统分析与综合。特别地，在随机逼近和随机优化中，推导出带有风险评估的严格停止规则。在鲁棒约束和平均约束条件下，提出了一种新的场景优化方法。然后，建立了考虑噪声输入输出数据的模型失效，并通过随机化算法考虑了变采样率控制系统的综合。

Analysis and synthesis of control systems were investigated via risk based optimization from both theoretical and application viewpoints. In particular, rigorous stopping rules with risk assessment were derived for sequential randomized algorithms in stochastic approximation and optimization. A new scenario approach was also developed for optimization subject to robust and average constraints. Then, model invalidation with respect to noisy input-output data was established, and synthesis of control systems with variable sampling rate was considered via randomized algorithms.