Canonical Duality Approach for Nonlinear Dynamical Systems

Abstract: This paper presents a canonical dual approach for solving a nonlinear population growth problem governed by the well-known logistic equation. Using the finite difference and least squares methods, the nonlinear differential equation is first formulated as a nonconvex optimization problem with unknown parameters. We then prove that by the canonical duality theory, this nonconvex problem is equivalent to a concave maximization problem over a convex feasible space, which can be solved easily to obtain global optimal solution to this challenging problem. Several illustrative examples are presented.