Title: Distributed Control of Positive Systems

Abstract: It is shown that several classical open problems in linear system theory, such as computation of stabilizing static output feedback gains and optimization of decentralized controllers, can be readily solved for positive systems using linear programming. In particular, optimal solutions can be verified for large-scale systems using computations that scale linearly with the number of interconnections. The restriction to positive state space models is natural in many applications, including stochastic systems, transportation networks and economics. Moreover, the main results are generalized to frequency domain representations using the notion of positively dominated system.