Abstract. The algorithm of finding all shortest paths in undirected network is considered. Two criteria are used: the minimum number of arcs in the path and minimum path length. The algorithm is analyzed for complexity and it is empirically shown that as the network density increases, the computational efficiency of the proposed algorithm becomes higher than that of the Floyd algorithm adequately modified to find the shortest path by two criteria.

Keywords: multicriteria problem of finding shortest paths, algorithm, computational complexity.