Improvement of Dinic algorithm based on ISAP algorithm

ISAP algorithm improvements to the DINIC algorithm:
In the preface to the Rujia graph theory, it is pointed out that the optimization of the algorithm itself is more complicated, and these high-quality graph theory algorithms are the wisdom crystallization of countless excellent algorithm designers.
If 1:30 will not understand clearly, it is also normal. But for a good acmer, the algorithm itself, can exercise your thinking. Gain Insight!

Here's what I know about Dinic and ISAP:

Dinic algorithm compared to the value of the EK algorithm, has been greatly improved, its advantages where? Is in his layered mind. Augmented on a hierarchical chart. However, he also has drawbacks.
That is, each time the augmentation is done, the hierarchy diagram is started again.
The dinic algorithm has the biggest advantage is the concept is simple, and the speed is good!!!

If the efficiency requirements are high, the dinic algorithm can be rewritten as an iterative form. But generally do not do so, but the use of ISAP algorithm;

The ISAP algorithm does the article on the level diagram just now:
1, first of all, the definition of the level diagram has modified D (i) to indicate the lower bound of the distance to the meeting point, along the feasible flow, when we can not find the augmented path, in the Dinic algorithm, is a one-time modification of all
The distance designator, while the ISAP is only modified by the augmented edge. This is the first optimization. Note that if the augmented path is not found from node I, D (i) >=n;
2, Gap Optimization: For each distance designator D (i) =x; with a num[x] array maintenance, that is, the number of x this distance designator, if a distance designator num[x] = 0;
There must be no augmented road.

Improvement of Dinic algorithm based on ISAP algorithm