Capacities of undirected cuts #
This file defines the capacity of a cut in a weighted multigraph. Parallel edges contribute separately and loops never contribute. Aggregating the weights of all edges with the same unordered pair of distinct ends therefore preserves every cut capacity.
For a simple graph, the multigraph cut capacity of Graph.ofSimpleGraph G agrees with
TauCeti.pairCutCapacity when pair capacities vanish away from the edges of G.
Main definitions #
Graph.edgeCutCapacity: the capacity of a cut in a weighted multigraph.Graph.pairCapacity: aggregation of multigraph edge weights by their distinct ends.SimpleGraph.ofSimpleGraphCapacity: restriction of pair capacities to actual edges.
Main result #
Graph.edgeCutCapacity_one: a unit-capacity cut has value equal to the number of crossing edge identities.Graph.pairCutCapacity_pairCapacity: aggregation preserves every cut capacity.Graph.isSubmodular_edgeCutCapacity: an edge-capacity function nonnegative on nonloop edges has a submodular cut function.Graph.isSymmSubmodular_edgeCutCapacity: the bundled symmetric-submodular result.SimpleGraph.edgeCutCapacity_eq_pairCutCapacity: the simple-graph and multigraph cut-capacity interfaces agree.
The defining finite-sum formula for edgeCutCapacity.
Giving every actual edge capacity one makes the capacity of a cut equal to the number of crossing edge identities. In particular, parallel edges are counted separately.
Aggregate edge weights by the unordered pair of their distinct ends. The capacity of every diagonal pair is zero, so loops are discarded.
Instances For
The defining finite-sum formula for pairCapacity.
On a pair of distinct vertices, pairCapacity sums exactly the weights of edges linking
those vertices.
Aggregating parallel edge weights by their distinct ends preserves every cut capacity.
Multigraph cut capacity is unchanged when the two sides of the cut are exchanged.
The cut function of edge weights nonnegative on nonloop edges is submodular. Together with
edgeCutCapacity_compl, this makes it a symmetric submodular set function. Parallel edges are
counted separately, and loop weights are unrestricted.
The cut function of edge weights nonnegative on nonloop edges is symmetric submodular. Loop weights are unrestricted.
The vertex set of the multigraph realization of a finite simple graph is finite.
The edge set of the multigraph realization of a finite simple graph is finite.
Restrict capacities on unordered vertex pairs to the actual edges of
Graph.ofSimpleGraph G.
Instances For
Include a vertex of a simple graph in the full vertex set of Graph.ofSimpleGraph G.
Instances For
Include a finite vertex set in the full vertex set of Graph.ofSimpleGraph G.
Instances For
The multigraph cut capacity of Graph.ofSimpleGraph G agrees with pair cut capacity when
capacities vanish away from the edges of G.