Numerical methods for nonlinear optimal control problems: Application to abnormal problems (Q806971)

One considers the standard optimal control problem which consists of maximizing the integral of a nonlinear function subject to nonlinear dynamics. By using various contrivances, this problem is converted into a variational one, somewhat equivalent to the determination of the trajectory of a system in analytical mechanics. Then one gets a block diagonal system of equations which is solved in practice by Newton's method. The technique can incorporate equality and equality constraints.