Dissipativity-augmented multiobjective control of networks
Ethan J. LoCicero, Leila Jasmine Bridgeman · International Journal of Control · 2025
This work is motivated by large-scale, interconnected systems with heterogeneous nonlinearities and objectives. A decentralised controller is designed to guarantee robust L2 stability for a network with uncertain, delayed, and nonlinear subsystems while optimising multiple nominal objectives: H∞, coloured-H2, and pole placement. To achieve this, common performance objectives and plant uncertainty characterisations are surveyed, and this paper shows how they can be united with the Dissipativity Theorem and modest generalisations of related results. The design problem is NP hard due to bilinear and trilinear constraints. A feasible point is constructed under mild conditions, and the convex-concave procedure locally minimises the cost function. Two numerical examples are provided.