The Art Gallery Guardian

Bounds on number of cuts


Consider we have an undirected graph of vertices, and there are positive cost on the edges. We define , to be the cost (value) of .

Let be a partition of where each partition class is non-empty. We define to be the set of edges with end points in two different partition classes. A set of edges is called a -cut, if for some such that .

We stress that by this definition, a -cut is always a -cut for . A cut is defined as a -cut. A min--cut is a -cut of minimum value (cost). We let to denote the value of the min--cut. A -approximate -cut is a cut of value at most .

It is well known that the number of min-cuts in a graph is . [1]

Unless specifically stated, we assume is a fixed integer at least , and is a fixed value at least .

We can express the state alternatively.

Theorem1

The number of cuts such that is .

1 Bounds related to approximate min-cuts

What happens when we want to know about the number of cuts with value at most ?

By simply analyzing Karger’s algorithm, one can obtain the following.

Theorem2

The number of cuts such that is .

With more careful analysis using tree packing, Karger added a floor function to the exponent [2].

Theorem3

The number of cuts such that is .

Indeed, we do have a lower bound of . Just consider an unweighted cycle, where min-cut has value . We can pick any edges, which forms a cut of value at most times the min-cut.

Note that we require to a fixed value. This is because there is a dependency on hidden inside the big . Our lower bound is absolute and does not depending on being fixed. It would be interesting to obtain a matching upper bound for arbitrary . Hence we can consider the problem with strict inequality.

Conjecture4

The number of cuts such that is .

Henzinger and Williamson showed the conjecture true for all [3].

2 Bounds related to -approximate min--cut

There are multiple ways to obtain the following theorem. For example, directly generalize Karger’s argument for cut counting.

Theorem5

The number of cuts such that is .

One can show a lower bounds of the form . Again, a cycle would be an example of such lower bound. The min--cut has value , and can be obtained by picking any edges. Recently, the upper bound has been closed (up to constant factor if is fixed) [4].

Theorem6

The number of cuts such that is . Namely for fixed .

In fact, they proved it by prorving the stronger theorem on -approximate min--cuts.

Theorem7

The number of cuts such that is . Namely for fixed .

Unlike -approximate min-cuts, there is no floor function in the exponent. Hence a natural conjecture would be.

Conjecture8

The number of cuts such that is .

3 Bounds on (approximate) parametric cuts

Consider we have weight functions . Define for . We are interested in knowing about cuts such that is bounded by , where is the min--cut value when the cost function is .

Karger showed the following [5].

Theorem9

The number of cuts such that for some is .

A even more general theorem follows, allowing -approximate -cuts.

Theorem10

The number of cuts such that for some is .

Of course, knowing the current result for number of -approximate -cuts, we would conjecture the following for parametric cuts.

Conjecture11

The number of cuts such that for is .

Note, we might relax the requirement that all and are non-negative. Aissi et. al. showed the following [6], only assuming .

Theorem12

The number of cuts such that for some is .

Can we obtain a stronger bound for ?

Parametric min-cut is related to multicriteria min-cut, which also have a lot of open problems [7].

4 Projected cut bounds?

Let , i.e. the minimum value of a cut containing . Fung et. al. showed a projected generalization of the cut counting bound [8] by ignore edges appeared in small cuts. Let .

Theorem13

The number of sets of the form where is a cut such that is .

If we let be the min-cut value, this is precisely the -approximate cut counting bound.

We can ask if all our theorem can have a projected cut version. We don’t know if it extends to -cuts (or what is the correct generalization).

5 What about hypergraphs?

Most of the above results in graphs generalizes to rank hypergraphs with an extra factor related to , often is . See the table in [7] for a survey. However, almost all exact min-cut counting problem are unknown for hypergraphs with unbounded rank.

Conjecture14

All the exact min-cut counting can be generalized to hypergraphs with unbounded rank.

One can construct a hypergraph where every set induces a distinct -approximate min-cut for all , so counting -approximate min-cut is unintresting for hypergraphs of unbounded rank.

References

[1]
D. Karger, C. Stein, A new approach to the minimum cut problem, Journal of ACM. 43 (1996) 601–640.
[2]
D.R. Karger, Minimum cuts in near-linear time, J. ACM. 47 (2000) 46–76 10.1145/331605.331608.
[3]
M. Henzinger, D.P. Williamson, On the number of small cuts in a graph, Information Processing Letters. 59 (1996) 41–44 https://doi.org/10.1016/0020-0190(96)00079-8.
[4]
A. Gupta, D.G. Harris, E. Lee, J. Li, Optimal bounds for the k-cut problem, J. ACM. 69 (2021) 10.1145/3478018.
[5]
D.R. Karger, Enumerating parametric global minimum cuts by random interleaving, in: Proceedings of the Forty-Eighth Annual ACM Symposium on Theory of Computing, ACM, New York, NY, USA, 2016: pp. 542–555 10.1145/2897518.2897578.
[6]
H. Aissi, A.R. Mahjoub, S.T. McCormick, M. Queyranne, Strongly polynomial bounds for multiobjective and parametric global minimum cuts in graphs and hypergraphs, Mathematical Programming. 154 (2015) 3–28 10.1007/s10107-015-0944-8.
[7]
C. Beideman, K. Chandrasekaran, C. Xu, Multicriteria cuts and size-constrained k-cuts in hypergraphs, Mathematical Programming. 197 (2023) 27–69 10.1007/s10107-021-01732-0.
[8]
W.-Shing. Fung, Ramesh. Hariharan, N.J.A. Harvey, Debmalya. Panigrahi, A general framework for graph sparsification, SIAM Journal on Computing. 48 (2019) 1196–1223 10.1137/16M1091666.
Posted by Chao Xu on .
Tags: Graph Theory.