A solution technique for the nonlinear inverse problem is the method of successive approximations in policy space. This technique is based upon gradients in which the approach to an optimal policy is by successive steps. The method of derivation employs the familiar concepts and techniques of dynamic programming. We essentially guess a presumably nonoptimal decision sequence. By simple reasoning we derive a set of recurrence relations that can be used to evaluate the effect of a small change in the decision sequence. We then use this information about the effect of decision changes to generate a new, improved, sequence of decisions. The effect of changes in the new sequence is evaluated. This iterative process is continued until no further improvement is possible. Each successive solution we obtain will be feasible for the problem, but not optimal.