Adjoin the dynamical
constraints to J using Lagrange multipliers l
Setting ¦F/¦xi = 0 gives an equation for li, the adjoint of xi:
The adjoints to the control variables are given by ¦F/¦ui and ¦F/¦xoo as:
The gradient vector is given by ¦F/¦ui and ¦F/¦xoo, so the optimal u and xo may then be found
using oneÕs favorite descent
method.
I use the BFGS method in order to obtain an estimate of the leading terms in the covariance matrix.