“…In a series of papers (see, e.g., [43,48,[59][60][61][62][63][64]), various authors tried to reduce several of the non-intrinsic drawbacks of this method, for example, by combining the obtained Zames-Falb multipliers with those corresponding to the Popov criterion or by allowing for higher order approximations. Still, for k-fold repeated nonlinearities, numerical tractability suffers because a non-convex search must be performed for a k-dimensional parameter [25,44,65]; also, the use of a common Lyapunov function for all L 1 -constraints and stability LMIs might cause additional conservatism for increasing values of k. Our approach avoids these troubles and is guaranteed to achieve no worse results, because it freely combines multiple pole Zames-Falb multipliers (each based on asymptotically exact parameterizations) with those from other criteria for repeated nonlinearities.…”