Abstract
This paper addresses the problem of dynamic input allocation in the presence of plant uncertainties. The current state of the art shows how to design an Allocator as the cascade of an Optimizer and an Annihilator to achieve steady-state input optimality and output invisibility simultaneously. This work proposes a novel algorithm based on polynomial factorization to design a dynamic Annihilator. The critical aspect of this approach lies in the assumption of the perfect plant knowledge, making the Annihilator not robust to uncertainties. A robustification process is introduced by optimizing its design parameters. This approach is formulated as a model-matching problem aiming to reduce the output mismatch induced by the allocation scheme while maintaining steady-state optimality. As the numerical simulations highlight, this method applies to linear and nonlinear allocation problems.
