EN
We optimize transition rules of bonus-malus system to achieve possibly best premium elasticity as defined by Loimaranta [1972] and later generalized as global elasticity by De Pril (1978). We use premium scale given by Norberg (1976). This issue constitutes a nonlinear nonconvex discrete optimization problem. To solve this problem, we apply improved greedy optimization algorithm, similar to one proposed by Morlock (1984). We analyse systems of different size for portfolios characterized by inverse Gaussian risk structure function with various parameters. We also propose alternative measures of global elasticity.