Distributed nonlinear optimal control using sequential convex programming and smoothing techniques

We regard a network of coupled nonlinear dynamical systems that we want to control optimally. The cost function is assumed to be separable and convex. The algorithm we propose to address the numerical solution of this problem is based on two ingredients: first, we exploit the convex problem structure using a sequential convex programming framework that linearizes the nonlinear dynamics in each iteration. Second, we use distributed dual decomposition methods to address the decomposable convex subproblems, that allow efficient parallel implementation. We analyze the convergence of the algorithm towards a local solution.