衡量连通图连通性一些指标(r-reachable, r-robust)

r-reachable

introduction

Definition 1(r-reachable~set): For any given $r\in \mathbb{N}$, a subset of nodes $S\subset N$ is said to be r-reachable if there exists a node $v_i\in \mathcal{V}$ such that $\Vert {\mathcal{N}}_i\backslash S\Vert \le r$. $r$-reachable set is a tool to measure the connectivity among the graph, the bigger $r$ is, the slighter one node be influenced by one of its neighbor. For example, a complete connected graph with $m$ nodes has $m$-reachable set. Hence, the influence of a regular nodes receiving adversary message by a malicious node drops to minimize.

code(Matlab)

function [result] = r_reachable( graph, subset)
temp = 0;
len = length(subset);
gsize = size(graph, 1);
result = 0;
for i = 1:len
    for j = 1:gsize
        if(subset(i)~=j && ~ismember(j,subset) && graph(subset(i),j)~=0)
            temp = temp+1;
        end
        continue
    end
    if result < temp
        result = temp;
    end
    temp = 0;
end

end

r-robust

introduction

Definition 2(r-robust~graph): For $r\in \mathbb{N}$, a graph $G$ is said to be $r$-robust if for all pairs of disjoint nonempty subsets $S_1,S_2\subset \mathcal{V}$, at least one of $S_1$ or $S_2$ is $r$-reachable. The usage of $r$-robust graph is similar to $r$-reachable set. The distinguish between is that $r$ -reachable represents local connectivity while $r$-robust represents global connectivity of a graph.Codes is given in index, significantly, it’s really complex to judge a graph is a r-robust with some r. In my opinion, it’s not a good index to measure a rather huge network. For example, we consider a network with 8 nodes, we need to split subset from 8 nodes totally 162 times.($C_8^1+C_8^2+C_8^3+C_8^4=162$)

code(Matlab)

function [result] = r_robust( graph )
gsize = size(graph, 1);
temp1 = 0;
temp2 = 0;
result = gsize;
for i = 1:floor(gsize/2)
    combine = nchoosek([1:gsize], i);   
    for j = 1:size(combine, 1)   
        temp1 = r_reachable(graph, combine(j,:));
        corresp = [];
        for k = 1:gsize
            if(~ismember(k, combine(j,:)))
                corresp(end+1) = k;
            end
        end
        temp2 = r_reachable(graph, corresp);
        if(temp1 > temp2)
            temp2 = temp1;
        end
        if(temp2 < result)
            result = temp2;
        end
    end
end
end