Closed loop control design for nonlinear, nonquadratic systems

The problem of optimal control of a nonlinear system with non quadratic final cost function admits a closed-loop suboptimal solution by approximation of the Lyapounov function by means of the control-equivalent Gaussian sum method. Using the duality between optimal control for deterministic systems and estimation for stochastic systems, the solution is presented as the superposition of solutions of linear subsystems.