What is the most efficient way to recompute maximum flow in a graph when:
we increase flow on one edge by one
we decrease flow on one edge by one
In the first case, is is enough to run one iteration of Ford-Fulkerson algorithm? In the second case, we need to recompute maximum flow only if the edge is part of a set of edges of maximum flow. Is it also enough to run one iteration of Ford-Fulkerson?