Basic concept of graph
Undirected graph
Edges have no direction, that is, bidirectional
结点V = { v 1 , v 2 , . . . , v 7 } \mathcal{V} = \{ v_1,v_2,...,v_7\}V={ v1,v2,...,v7}
边ε = { and 1 , 2 , and 1 , 3 , . . . , e 6 , 7 } \varepsilon = \{e_{1,2},e_{1,3},...,e_{6,7}\}e={ e1,2,e1,3,...,e6,7}
图G = { V , ε } \mathcal{G} = \{ \mathcal{V},\varepsilon \}G={ V,e }
directed graph
Edges are directional, that is, unidirectional
Unweighted graph
The edge has no weight, which can also be understood as the weight is 1
entitled figure
Edges have weights
Graph Data Structure
Adjacency matrix and adjacency list
undirected unweighted graph
Adjacency list
Indicates who the node is connected to,
vertex | Neighbors |
---|---|
1 | 2,3,4 |
2 | 1,4 |
3 | 1,4,6 |
4 | 1,2,3 |
5 | empty |
6 | 3,7 |
7 | 6 |
Adjacency matrix
[ 0 1 1 1 0 0 0 1 0 0 1 0 0 0 1 0 0 1 0 1 0 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 1 0 0 0 0 0 1 0 ] \begin{bmatrix} 0 & 1& 1& 1& 0& 0& 0\\ 1 & 0& 0& 1& 0& 0& 0\\ 1 & 0& 0& 1& 0& 1& 0\\ 1 & 1& 1& 0& 0& 0& 0\\ 0 & 0& 0& 0& 0& 0& 0\\ 0 & 0& 1& 0& 0& 0& 1\\ 0 & 0& 0& 0& 0& 1& 0 \end{bmatrix} 0111000100100010010101110000000000000100010000010
The adjacency matrix represents the nodes iii and nodejjThe weight between j , if it is an unweighted graph, the weight is 1. The adjacency matrices of undirected graphs are all symmetric. Because edges are bidirectional.
directed unweighted graph
Adjacency list
Indicates who the node is connected to,
vertex | Neighbors |
---|---|
1 | 2,4 |
2 | 4,5 |
3 | 1,6 |
4 | 5 |
5 | empty |
6 | 7 |
7 | 6 |
Adjacency matrix
[ 0 1 0 1 0 0 0 0 0 0 1 1 0 0 1 0 0 0 0 1 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 1 0 ] \begin{bmatrix} 0 & 1& 0& 1& 0& 0& 0\\ 0 & 0& 0& 1& 1& 0& 0\\ 1 & 0& 0& 0& 0& 1& 0\\ 0 & 0& 0& 0& 0& 1& 0\\ 0 & 0& 0& 0& 0& 0& 0\\ 0 & 0& 0& 0& 0& 0& 1\\ 0 & 0& 0& 0& 0& 1& 0 \end{bmatrix} 0010000100000000000001100000010000000110010000010
The adjacency matrix represents the nodes iii and nodejjThe weight between j . The list shows the starting point, and the row shows the reached point, such asa 1 , 2 = 1 a_{1,2} = 1a1,2=1 means there is a unidirectional edge from node 1 to node 2.
entitled figure
Adjacency matrix
[ 0 2 4 1 0 0 0 2 0 0 3 0 0 0 4 0 2 0 0 5 0 1 3 2 0 0 1 0 0 0 0 0 0 0 0 0 0 5 0 0 0 1 0 0 0 0 0 1 0 ] \begin{bmatrix} 0 & 2& 4& 1& 0& 0& 0\\ 2 & 0& 0& 3& 0& 0& 0\\ 4 & 0& 2& 0& 0& 5& 0\\ 1 & 3& 2& 0& 0& 1& 0\\ 0 & 0& 0& 0& 0& 0& 0\\ 0 & 0& 5& 0& 0& 0& 1\\ 0 & 0& 0& 0& 0& 1& 0 \end{bmatrix} 0241000200300040220501300000000000000510010000010
In fact, it is very similar to an undirected graph, which means the link between two nodes. The weighted graph has weights, but the unweighted graph does not.
degree matrix
The degree matrix is an important concept in graph theory, which is used to describe the degree of each node in an undirected or directed graph. It is a diagonal matrix, the elements di d_i in the matrixdirepresents node iidegree of i , i.e. with nodeiii is the number of connected edges.
Undirected graph
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 ] \begin{bmatrix} 3 & 0& 0& 0& 0& 0& 0\\ 0 & 2& 0& 0& 0& 0& 0\\ 0 & 0& 3& 0& 0& 0& 0\\ 0 & 0& 0& 3& 0& 0& 0 \\ 0 & 0& 0& 0& 0& 0& 0\\ 0 & 0& 0& 0& 0& 2& 0\\ 0 & 0& 0& 0& 0& 0& 1 \end{bmatrix}
3000000020000000300000003000000000000000200000001
It can be seen that is a diagonal matrix.
directed graph
For directed graphs, it can be divided into out-degree matrix and in-degree matrix. The out-degree matrix represents the out-degree of each node, and the in-degree matrix represents the in-degree of each node. This is easy to calculate, not in the demonstration
Laplacian matrix
The Laplacian matrix is used to describe the properties of graphs. There are generally several methods for calculating the Laplacian matrix
1. Symmetric normalized Laplacian: Let L be the Laplacian matrix of graph G, D be the degree matrix, then the symmetric normalized Laplacian matrix is L sym = D − 1 2 LD − 1 2 L_{sym}= D^{-\frac{1}{2}}LD^{-\frac{1}{2}}Lsym=D−21LD−21。
2、Random walk Laplacian:设 M = D − 1 L M=D^{-1}L M=D− 1 L, then the random walk Laplacian matrix isL rwl = I − M L_{rwl} = IMLr wl=I−M ,III is the identity matrix.
3. Unnormalized Laplacian: The Laplacian matrix without normalization can be expressed as L = D − AL=DAL=D−A , where A is the adjacency matrix and D is the degree matrix
Matrix properties[1]
-
0 is his eigenvalue, 1 N 1_N1Nis the eigenvector
-
x T ℓ x = 1 2 ∑ i = 1 N ∑ j = 1 N a i j ( x j − x i ) 2 x^T\ell x=\frac{1}{2} \sum_{i=1}^{N} \sum_{j=1}^{N} a_{ij}(x_j-x_i)^2 xTℓx=21∑i=1N∑j=1Naij(xj−xi)2,ℓ \heThe positive semi-definiteness of ℓ means that ℓ \ellAll characteristic roots of ℓ are real and nonnegative
-
If G is connected, then ℓ \ellThe second smallest characteristic root of ℓ , you can use λ 2 ( ℓ ) \lambda_2(\ell)l2( ℓ ) , called the algebraic connectivity of G, which is greater than 0
-
The algebraic connectivity of G is equal to minx ≠ 0 , 1 NT x = 0 x T ℓ xx T x min_{x \ne 0,1_N^Tx=0} \frac{x^T \ell x}{x^T x}minx=0,1NTx=0xTxxTℓx, so if $1_N^T x = 0,x^T \ell x \ge \lambda_2(\ell)x^Tx $ .
references
[1] R. Olfati-Saber, R.M. Murray, Consensus problems in networks of agents with switching topology and time-delays. IEEE Trans. Autom. Control 49(9), 1520–1533 (2004)
The pictures in this article are from the course of Mr. Wang Shusen. The course has not been carefully studied, and only these basic knowledge are needed.