November 28, 2024973 words

图论

I am forced to study this goddamn course as a requirement, and I hate the outdated overly verbose nonsense textbook, but I need to do it. I need to pass this damn course.

I have literally been joyless these days-- very depressed. Anyways, I hate school, I hate everything.

Note: Graph defaults to undirected, simple graph

Chapter 1

  • 简单图: no loops, no more than one edge between any two vertices.
  • 完全图: a graph in which each vertex is connected to every other vertex.
  • 生成子图: subgraph that includes all the vertices of the original graph
  • 导出子图: induced subgraph (subgraph but you can't erase edges)
  • 行迹: a path with no redundant edge
  • 轨道: a path with no redundant vertex
  • 回路: a path with the same starting and ending point
  • 圈: cycle
  • 有向图: directed graph
  • Dijkstra: loop: find min dist unvisited and update dist for unvisited, O(|E|+|V|log|V|)
  • Bellman-Ford: (negative weight) loop: relax all edges (at most |V|-1 times), O(|V||E|)

Chapter 2

  • 树: connected, acyclic graph
  • 树叶: nodes where degree is 1
  • 分支点: nodes where degree is greater than 1
  • 森林: acyclic graph
  • 生成树: spanning tree
  • Kruskal: sort the edge, loop: add smallest edge in the graph(if acyclic), O(|E|log|V|)
  • Prim: basically Dijkstra (you can aggregate all paths after Dijkstra, very trivial), O(|E|+|V|log |V|)
  • Huffman 树: combing smallest node each time

Chapter 3

  • 割集:a partition of the vertices of a graph into two disjoint subsets
  • Menger Theorem
  • Whitney Theorem
  • k 扇形: k disjoint path from X to Y
  • k-connected-graph: has more than k vertices and remains connected whenever fewer than k vertices are removed
  • Dirac Theorem

Chapter 4

  • 平面图: Planar graph, a graph that can be embedded in the plane(<=> on a sphere) without any edges crossing
  • 桥: an edge if deleted, connected components increases
  • Thm(generalized Euler): V-E+F-C= 1, f: area, c: component, prove: induction (trivial)
  • Thm: sum(deg(F))= 2 * |E(G)| (trivial)
  • Corollary: In planar graph E <= 3V-6 (trivial)
  • Corollary: In planar graph min(degree(V))<= 5 (trivial)
  • K5: 5 vertices complete graph (WHY SO CONFUSING??????)
  • K3,3: complete bipartite graph with set of 3 vertices each (WHY SO CONFUSING??????)
  • 极大平面图: planar but adding any edge would not be planar
  • Thm: If V>3, then maximum planar graph <=> all faces are triangles (trivial) <=> E = 3V-6
  • 同胚: both can be obtained from the same graph by subdivisions of edges

Chapter 5

  • 匹配: set of edges, no two sharing a vertex
  • 完备匹配: all nodes are included
  • 最大匹配: match with most edges
  • 交错轨道: a path that alternates between edges in and not in the matching
  • 可增广轨道: a path that starts and ends on free vertices, and alternates between edges in and not in the matching
  • Berge: a matching M is maximum iff there is no augmenting path with M.
  • Hall: there is an X perfect matching iff for every subset W of X, |W| <= |NG(W)|
  • 覆盖: a set of vertices that includes at least one endpoint of every edge of the graph.
  • Kőnig
  • Tutte
  • Hungarian Algorithm

Chapter 6

  • Euler 迹: a path traversing every edge <=> at most 2 vertex degree is odd
  • Euler 回路: a closed path travering very edge <=> no vertex degree is odd
  • Euler 图: a graph with a Eulerian cycle
  • Hamilton 轨道: a path traversing every vertex in the graph
  • Hamilton 圈: a cycle traversing every vertex in the graph
  • Hamilton 图: a graph with Hamilton cycle
  • Bondy–Chvátal theorem Theorem

Chapter 7

  • Thm: Bipartite graph with edges <=> Х(G)=2 (trivial)
  • k 边着色: color edges with k different colors
  • 正常 k 边着色: color edges with k different colors with different colors on adjacent edges
  • Lemma: If G isn't an odd cycle, then exist a 2-coloring of edges, st every degree(vertex)>=2, 2 colors appear in the edges connected
  • 改进:edge coloring using fewer colors
  • 最佳边着色:edge coloring using fewest colors
  • Δ: max(degree(vertex))
  • Vizing: Graph minimum edge coloring is max(degree(vertex)) colors or max(degree(vertex))+1 colors, Proof: https://en.wikipedia.org/wiki/Vizing's_theoremt
  • Planar graph five coloring: Induction -> '(if can't) one vertex degree=5 with different cyclic neighbors colors 1,2,3,4,5 -> 2 color subgraph G(1,3) (G(2,4)) with node 1,3 (node 2,4) on one path -> planarity contradiction
  • Planar graph four coloring: computer needed
  • 颜色多项式: Chromatic polynomial P(G, k) counts the number of its vertex at most k-colorings
  • Thm: P(G, K) = P(G-e, K)- P (G combine u,v, k) for every simple graph G and edge e=(u,v) (trivial)

Chapter 8

  • 底图: the undirected graph for directed graph
  • 外邻定点: u->v, v is the external point
  • 内邻集: all internal points
  • 外邻集: all external points
  • 强连通: u can go to v AND v can go to u
  • 单项连通: either u can go to v or v can go to u (or both)
  • 弱连通: the undirected path can reach each other
  • 竞赛图: directed, complete graph
  • Thm: Every k-chromatic graph has a k-chromatic subgraph of min degree at least k−1 (trivial)
  • Thm: Every directed, complete graph is hamiltonian (trivial)
  • Thm: Length of longest path in a directed graph is greater than smallest degrees inbound (and outbound) in all vertex (trivial)
  • Thm: In a directed graph if minimal inbound and outbound of every vertex in a raph is greater or equal than vertex/2, it is hamiltonian (trivial)

Chapter 9

Loading...




Loading...