[en] An automatic tool, INERLOAD, is developed to search for the optimal in-core fuel loading pattern of boiling water reactors by using genetic algorithm. Sub-pattern swap concept is applied for the crossover operator. Different fresh fuel check-board arrangements including two standard checkerboard patterns and experience-based check-board patterns are implemented into the code. The objective function of a loading pattern consists of the k_eff at the end of cycle, the maximum power peaking factor, and the cold shutdown margin at the beginning of cycle with the corresponding weighting factors. In order to evaluate these parameters, the steady-state core simulator, SIMULATE-3, is utilized to perform the three-dimensional calculations. Haling axial power shape is applied to estimate k_eff at the end of cycle and maximum power peaking factor while using SIMULATE-3. The code is tested for both long and short cycle length cases which are based on an actual core of BWR-6 reactor, Kuosheng Nuclear Power Plant. Results show that this automatic tool is very competitive to well-trained engineer's designs for both cases. (authors)