Dynamic norms for random and directed mutations in a genetic algorithm for solving the problem of placing elements in an orthonormal lattice have been developed. The principle of these norms is borrowed from the method of imitation annealing. The probability of using random mutations decreases in accordance with the annealing temperature, and the probability of directional mutations increases from iteration to iteration when approaching the optimum.
Keywords: genetic algorithm, the task of placing the elements, the annealing method, random mutations, directed deterministic mutations