Author: Stefan Richter 2021-08-21 10:12:50
Published on: 2021-08-21T10:12:50+00:00
In this context, Stefan compliments Zmn on his lateral thinking and mentions that he will need to give it a closer read. Stefan believes that the computeCost definition is important in solving min-cost flow problems and notes that it generally has to be separable and convex. He thinks that any cost function that implies a fixed cost per edge is concave and problematic for min-cost flow algorithms because it could reduce some kind of NP-hard selection problem to it. However, he acknowledges that there are tools like approximation and parameterized algorithms as well as heuristics to address this issue. ZmnSCPxj responds by questioning why cost should be a number and suggests exploring alternative structures that can provide the necessary operations demanded by the mincostflow algorithm.
Updated on: 2023-06-03T05:34:32.278186+00:00