Nowadays public transit network design is highly urgent challenge in a field of road network analysis. The issue is known to be complex and cumbersome because of a huge number of factors that should be taken into consideration to cope with it. Transit network design is not limited just by topological optimization but requires consideration of such non-network parameters as transit vehicles fleet, frequency-settings, transit vehicle classes, etc. No wonder that the corresponding optimization problems appears to be NP-hard. Therefore available today approaches for transit network design handling actual-size road networks are all based on evolution or genetic algorithms. This paper is devoted to the case of public transit network design in real road network with 1280 actual bus stops. Methodological tools exploited to solve this problem are discussed, computational results are given.