Exact Solution Method for Robust Semidefinite Programming Problemsand Its Application to Control System Synthesis

在这项研究中，我们讨论了当我们分析受参数不确定性影响的控制系统的鲁棒性能时自然出现的鲁棒半定规划问题(Robust SDP)。对于稳健的SDP，我们成功地构造了一个保证渐近精确的SDP序列。另一方面，为了获得更好的不确定离散时间系统的鲁棒性能，我们提出了一种新的内存型周期时变控制器结构，并建立了利用SPS设计这种控制器的具体方法。江原义雄博士因其关于不确定离散时间系统的稳健性能分析的论文而获得SICE年度会议国际奖。

In this study we dealt with robust semidefinite programming problems (robust SDPs) that naturally appear when we analyze robust performance of control systems affected by parametric uncertainties. For the robust SDPs, we successfully constructed a sequence of SDPs that is ensured to be asymptotically exact. On the other hand, to achieve better robust performance for uncertain discrete-time systems, we conceived a novel memory-type periodically time-varying controller structure and established a concrete way to design such controllers by means of SDPs. Dr. Yoshio Ebihara received SICE Annual Conference International Award for his paper dealing with robust performance analysis of uncertain discrete-time systems.