Application of nonlinear optimization to reactor core fuel reloading

We use mixed-integer nonlinear programming (MINLP) as the optimization method for determining optimal loading schemes in nuclear reactor fuel management. Contrary to most current methods, which treat the physical core description as a black box, the MINLP approach uses the algebraic equations describing the physical core behavior directly in the optimization program. For a simple but still quite realistic model, we compare the obtained solution quality, the speed and the flexibility of this method with the solutions obtained by currently used optimization methods. It is shown that MINLP is competitive with current methods, and that MINLP combined with local search heuristics is a promising approach to fuel management optimization.