In this case the "greedy" algorithm was guaranteed to give the best possible flow in the network so this speedup applies to the optimal solution. I think in a l

In this case the "greedy" algorithm was guaranteed to give the best possible flow in the network so this speedup applies to the optimal solution. I think in a lot of cases greedy algorithms are used when the alternative is an NP-hard problem like bin packing or TSP for which we might be looking at speeding up an already sub-optimal solution.