Application of Bernstein expansion to the solution of control problems (Q1582464): Difference between revisions

The author provides an informative survey of the use of Bernstein expansion in control theory. In particular, he describes reliable methods for solving stability problems. Consider a family of polynomials \[ f(s,p)= a_0(p) s^m+\cdots+ a_{m-1}(p) s+ a_m(p), \] where the coefficients \(a_k(p)\) depend polynomially on the parameter vector \(p= (p_1,p_2,\dots, p_n)\). The parameter vector \(p\) is known to belong to a given box \(B\subseteq \mathbb{R}^n\). For \(p\in B\), the polynomial \(f(.,p)\) is \({\mathcal D}\)-stable, if all zeros of \(f(.,p)\) are inside \({\mathcal D}\), where \({\mathcal D}\) is a subset of the complex plane. The family is robust \({\mathcal D}\)-stable, if \(f(.,p)\) is \({\mathcal D}\)-stable for every \(p\in B\). Special choices of \({\mathcal D}\) lead to Hurwitz-stability, to Schur-stability, and to damping. Also covered is a related topic: Solving systems of strict polynomial inequalities over a given box \(B\subseteq \mathbb{R}^n\) with Bernstein expansion. Numerical examples using interval arithmetic illustrate the described methods.