Suggestions about D´Esopo-Pape algorithm and SPFA

[prprprpony]

Firstly, in the articles of Minimum-cost flow and assignment problem, D´Esopo-Pape algorithm is chose as the algorithm for shortest path.
However, its worst case time complexity is exponential.

Why not use SPFA instead?
Its implementation is basically the same as D´Esopo-Pape algorithm, but has better worst case time complexity (O(nm)).

Secondly, both SPFA and D´Esopo-Pape algorithm are said to be "quite fast in practice" in the ariticles, but none of them contain any figure from experiments to support the statement.

[jakobkogler]

I agree. I've never understood why D`Esopo-Pape is so much used on the original e-maxx. Especially since nobody else uses that one.

I'm not sure if I have time in the next days to replace implementations with SPFA. But you are welcome to make any pull-requests yourself.