OAI/Combinatorics/ContingencyTables/Transport/PublishedEdmondsKarp.lean

GitHub source file

10 declarations, 0 contain sorry, 0 declared as axioms

Shortest-augmenting-path bound

The finite integral-network specialization of Edmonds and Karp, Theoretical Improvements in Algorithmic Efficiency for Network Flow Problems, J. ACM 19(2) (1972), 248–264, Theorem 1 (printed page 251). The bound (n³−n)/4 implies the n³ bound for shortest residual augmentations.

Primary text: https://web.eecs.umich.edu/~pettie/matching/Edmonds-Karp-network-flow.pdf

  1. abbrev OAI.ContingencyTables.PublishedEdmondsKarp.Flow line 24
  2. abbrev OAI.ContingencyTables.PublishedEdmondsKarp.Capacity line 25
  3. def OAI.ContingencyTables.PublishedEdmondsKarp.Feasible line 29

    Source documentation

    Skew flows combine the two antiparallel directions. The upper capacity bound and skew symmetry give both directed residual bounds.

    def Feasible {n : ℕ} (c : Capacity n) (s t : Fin n) (f : Flow n) : Prop :=
    …

    Mentions: OAI.ContingencyTables.PublishedEdmondsKarp.Capacity, OAI.ContingencyTables.PublishedEdmondsKarp.Flow

  4. def OAI.ContingencyTables.PublishedEdmondsKarp.value line 34
  5. def OAI.ContingencyTables.PublishedEdmondsKarp.IsMaximum line 36
  6. def OAI.ContingencyTables.PublishedEdmondsKarp.ResidualPath line 40

    Source documentation

    A directed residual path with k arcs and no repeated vertices.

    def ResidualPath {n : ℕ} (c : Capacity n) (f : Flow n) (s t : Fin n)
        (k : ℕ) (p : Fin (k+1) → Fin n) : Prop :=
    …

    Mentions: OAI.ContingencyTables.PublishedEdmondsKarp.Capacity, OAI.ContingencyTables.PublishedEdmondsKarp.Flow

  7. def OAI.ContingencyTables.PublishedEdmondsKarp.ShortestAugmentation line 49

    Source documentation

    An actual saturating augmentation along a shortest residual path. The bottleneck is positive, bounded by every residual capacity, and attained by an arc. The displayed update adds it forward and subtracts it backward on precisely the traversed arcs.

    def ShortestAugmentation {n : ℕ} (c : Capacity n) (s t : Fin n)
        (f g : Flow n) : Prop :=
    …

    Mentions: OAI.ContingencyTables.PublishedEdmondsKarp.Capacity, OAI.ContingencyTables.PublishedEdmondsKarp.Flow

  8. def OAI.ContingencyTables.PublishedEdmondsKarp.AugmentationTrace line 63

    Source documentation

    A finite initial segment of an actual shortest-augmentation execution. It need not already have reached a maximum flow.

    def AugmentationTrace {n K : ℕ} (c : Capacity n) (s t : Fin n)
        (F : Fin (K+1) → Flow n) : Prop :=
    …

    Mentions: OAI.ContingencyTables.PublishedEdmondsKarp.Capacity, OAI.ContingencyTables.PublishedEdmondsKarp.Flow

  9. def OAI.ContingencyTables.PublishedEdmondsKarp.EdmondsKarp1972Theorem1 line 72

    Source documentation

    The shortest-augmenting-path bound for integral skew flows. Every finite initial segment has at most n³ augmentations, independently of the rule used to break ties between shortest paths.

    def EdmondsKarp1972Theorem1 : Prop :=
    …
  10. theorem OAI.ContingencyTables.PublishedEdmondsKarp.no_long_trace line 78