Core loading pattern optimization for research reactors

The algorithm based on successive mixed-integer linear programming and backward diffusion calculation is developed to determine the best locations of fuel assemblies in the core of research reactors. To determine optimum core loading patterns, the successive mixed-integer linear programming technique is applied. The backward diffusion calculation theory based on expression of fuel property in terms of power shape is used to lower computational cost of the algorithm by decreasing computer time necessary to calculate sensitivity coefficients. The algorithm developed is applied for real 2-D problems of the fuel management optimization for the VVR-M research reactor of the Kiev Institute for Nuclear Research (Ukraine).