is it possible that the minimum cost of spanning of a weighted graph is different by Kruskal's algorithm and by Prim's algorithm?