The design and operation of a nuclear reactor are posed as optimal control problems in terms of a generalized set of design objectives and a generalized control that influences the nodal material bucklings in a one-group spatially nodalized reactor model. The necessary conditions for optimality are derived by use of the Pontryagin Maximum Principle. An iterative algorithm is worked out for the resulting equations. A useful property of this algorithm is that each iteration produces an improved, consistent reactor life study for the assumed control. Therefore, the iterations may be terminated at any suboptimal yet acceptable stage. Furthermore, the designer may intervene in the iterative convergence toward the optimal control to exercise judgment and intuition not readily included in an algorithm. The approach is verified by solving a number of sample problems with the test code ØPTIM. The results of these problems show that the method works and quickly gives significant improvement in the design.