A simplified but still quite realistic model of the nuclear reactor fuel management problem is used to search for optimal fuel loading schemes. Several adaptations and model adjustments are described that make the model tractable for a general nonlinear mixed-integer solver. Results are compared with results from pairwise interchange optimization. Use of solutions from nonlinear mixed-integer optimization as starting values for local search heuristics leads to powerful optimization methods.
- nonlinear mixed-integer programming
- nonlinear assignment
- fuel management
- nuclear reactors