# 2020-03-20 Fundamental Concept I + Bonus Questions ###### tags: `圖論` [TOC] {%youtube 8XbaPEGTOvM %} ## 講義: fundamental_concept_0_2020 ### **Definition** * $V(G)$ : set of vertices * $E(G)$ : set of edges >[color=#00ff00] > > $E(G) ⊆ V(G) \times V(G)$ >[name=MathematiKai ] #### 2. relation 裡面的兩個元素是parrallize relation 是兩個集合之間cross product 的結果 map 到 0 或1 ex: 如果點集合是 $v_1,v_2 ..~ v_5$ 跟自己crossproduct 最多有25個情況 map 到 codomain 的值就只有 1 或 0, 有邊就是 1 反之亦然 graph 在定義並沒有限制 mutiple edge 的發生 所實際上要表示出 這三種關係 點的編號 邊的編號 和點點之間的關係式(只記下有關係 codomain = 1 的點) Graph 的兩個集合中 若是有mutiledge的存在,在edge的編號上做改變,在relation的edge集合中才會看出,相同element $(V_1,V_2)(V_1,V_2)$是指向不同的邊 無向圖 list 的 order 不重要 adjacent 是 討論兩個點之間的關係 有連線的兩個端點 是neighbor 專業講法是 adjacent ### Rooted tree vs. Unrooted tree Root: 點跟點之間,有 parent 和 child 關係 Unrooted: 反之亦然 Tree 是 bipartite 用著色的方法可以很直觀的證明 也可以用到root 的距離 是奇數還是偶數去分出兩個independent set Tree 有四種不同的定義(前三項任兩個合併就是tree) 1. connected + no cycles 2. connected + n-1 edges 3. n-1 edges + no cycles 4. no loops + exactly one path between vertex * **Path** : $<v_0,v_1...,v_k>$ a distinct adjacent vertices sequence * **Loop** : **an edge** whose endpoint are equal * **Cycle**: $<v_0,v_1...,v_k,v_0>$ $v_0\ ... v_k$ distinct sequence of adjacent vertices * **Multiple edges** : edges having the same pair of endpoints --- * **Simple graph** : graph having <font color='#ff0000' > no </font>loops or multiple edges * **Multigraph** : graph having loops or multiple edges * **Finite Graph** : $V(G)$ and $E(G)$ are finite sets * **Null Graph** : $V(G)$ and $E(G)$ are empty * **Complement Graph** : vertex Set: V(G) $\\uv \in E(\bar G)\Leftrightarrow uv \notin E(G)$ * **Subgraph :** $V(H) \subseteq V(G) \\ E(H) \subseteq E(G) \\點邊的關係相同\\Denoted \ as:H \subseteq G$ * **Connected graph :** G is connected $\Rightarrow$ each pair of vertices in G belong to a path --- * **Clique (點集合的子集)** : * **Definition**: a set of pairwise *adjacent* vertices * *Complete* (sub)graph * pairwise: 兩兩相聯都有邊, 單一一個點也算pairwise $\Rightarrow$ 1個點是clique * clique size 的最大size 是NP complete 問題 * **Independent set (stable set)** : * **Definition**: a set of pairwise *nonadjacent* vertices * Duel problem: $\text{finding Max clique in } (G) \Leftrightarrow \text{finding Max independent set in }(\bar G)$ ![](https://i.imgur.com/MOR70Gf.png =50%x) --- **為了方便記錄、去操作graph,需要用matrix的方式去紀錄圖** * **Adjancency Matrix(A)** : $a_{ij}$ is number of edges between $v_i, v_j$ (vertice-to-vertice) * **Incident Matrix(M)** : $m_{ij}$ is boolean of "$v_{i}$ is endpoint of $e_{j}$" (vertice-to-edge) - incident: 點為邊的端點的這種relation的專業講法, incident edges = degree of vertex ![](https://i.imgur.com/0C0uG4V.png =80%x) ### Königberg 7-Bridge Problem * 七橋問題是圖論的起源 * Situation: 每個edge都必須並只能走過1次 1. 每個點的degree都是偶數: 可從任意點開始,並全部走過且回到原點. 2. 只有2個點的degree是奇數: 例點A和點B, 那可以從其中一點(A或B)出發, 走到另一點(B或A). 3. 其餘無解 ### Partite * Independent set * Multipartite - union of independent sets + **Bipartite ⟺ No odd length cycle** + [證明P.28](https://hackmd.io/QIz8EtcKRqWANklnF4S-Fw?both) + k-partite: k $\geq3$ * Detect k of k-partite Graph - Color the Vertices * Complete k-partite graph $K_{n_1,n_2,...,n_k}$ * every vertex is connected with those in different partite ### Chromatic number $\chi(G)$ * Minimum number of color s.t. adjacent vertices' color are different. 一種把點做label的方式,adjacent vertex 不同顏色來著色,所需最少需要的顏色數 * $\chi (tree) =2$ $\\\chi(bipartite) \leq 2$ * $\chi$一定存在,但是如果用最少的color 去著色是無法決定難度的問題(harder than NP-C?) >[color=#00ff00] > > Chromatic number 有針對 vertex 與 edge 兩種 > 此處$\chi(G)$是針對vertex > > 找出$\chi(G)$等同找尋「k-partite的最小的k值」 > 應該是NP-C的問題 >[name=MathematiKai ] ### Isomorphism (圖跟圖之間的關係) * **Isomorphism** (General desciption) - set $X$ with a binary relation $R$ - set $Y$ with a binary relation $S$ - isomorphism from $X$ to $Y$ is a **bijective** function $f: X → Y$ such that $R(u,v) \Leftrightarrow S(f(u),f(v))$ * 圖論定義: * From a simple graph G to a simple graph H is a **bijective** function$f : V(G) \rightarrow V(H) \text{ such that } uv \in E(G) \Leftrightarrow f(u)f(v) \in E(H)$ where $u,v \in G$, <br/> denoted as $G\cong H$ * 較大的圖中找小圖的isomorphic 是NPC的問題 * 是一種equivalence 的關係 1. 自己跟自己同構(reflexive) 2. 滿足互為同構的關係(symmetric) 3. 同構擁有遞移性(transitive) * Proof: :::info <font color=#000000> 1. The identity permutation on $V(G)$ is an isomorphism from $G$ to itself. Thus $G \cong G$ 2. If $f : V(G) \rightarrow V(H)$ is $isomorphism$ from $G$ to $H$, <br/> by definition , $\exists f^{-1} : V(H) \rightarrow V(G)$ s.t. $uv \in E(H) \Leftrightarrow f^{-1}(u)f^{-1}(v) \in E(G)$ <br/> (bijective function 的特性)<br/>.Thus $G \cong H \rightarrow H \cong G$ 4. Suppose $f : V(F) \rightarrow V(G)$ and $g: V(G) \rightarrow V(H)$ are $isomorphisms$ s.t. <br/> $uv \in E(F) \Leftrightarrow f(u)f(v) \in E(G)$ <br/> $xy \in E(G) \Leftrightarrow g(x)g(y) \in E(H)$ <br/> since it's $isomorphism$, <br/>$\forall xy \in E(G),~ \exists uv \in E(F)$ s.t. $f(u)=x$, $f(v)=y$ <br/> thus yields $uv \in E(F) \Leftrightarrow g(f(u))g(f(v)) \in E(H)$ <br/> $F \cong G$, $G \cong H \rightarrow F \cong H$ </font> ::: ## Bonus Quetions 1. Prove "graph containing odd cycle" $\Leftrightarrow$ "graph is not bipartite" **Sol:** 提示 pigeon hole? 用著色的方式解好像挺容易的(? [name=Neko] 話說這題目有辦法從右邊到左邊? [name=Neko] $P \Leftarrow Q$ is $\neg P \Rightarrow \neg Q$ [name=Omnom] 3/27講義的第28頁 [ref1](https://proofwiki.org/wiki/Graph_is_Bipartite_iff_No_Odd_Cycles) 2. Define isomorphism of multiple graph **Sol:** [ref1](https://math.stackexchange.com/questions/3201645/expanding-definition-of-simple-graph-isomorphism-to-include-multigraphs)