Hi Everyone,
I'd like to create random adjacency matrices for networks in the following form (see image), as you can see I have 6 nodes (n = 6) and each node must have 2 connections (degree = 2) to two other nodes chosen randomly. So the sum of each row must equal 2(or whatever value "degree" is assigned). The matrix must be symmetric with zeros along the diagonal (properties of the adjacency matrix). I have written a code that generates random adjacency matrices but Imnot sure how to constraint the program such that each node has exaclty 2 random connections.
A = zeros(n,n); % Create initial matrix
for i = 1:n
for j = 1:n
if i==j
A(i,j)=0; % Zero values in the diagonal entries
else
A(i,j)=round(rand); % Choosing random 0 and 1 values
A(j,i)=A(i,j); % Make A matrix symmetric
end
end
end
Any help is greatly appreciated, Thank you!

 Réponse acceptée

David Hill
David Hill le 13 Mai 2021

0 votes

Graph has to be a polygon.
numNodes=6;
a=randperm(numNodes);
b=circshift(a,1);
g=graph(a,b);
plot(g);
m=full(adjacency(g));

3 commentaires

NeGM
NeGM le 13 Mai 2021
Hi David,
this is perfect given that each node has 2 connections but I'd like to know how to contraint it si that I can change the number of nodes and edges for instance how can I generate an adjacency matrix with n = 20 and degrees = 3 (number of connections) to 3 different nodes. any suggestions? Thank you so much in advance
David Hill
David Hill le 13 Mai 2021
Modifié(e) : David Hill le 13 Mai 2021
Some combinations of nodes and degrees will not work.
while 1
numNodes=20;
degrees=3;
n=repelem(1:numNodes,degrees);
s=[];t=[];
flag=1;
while ~isempty(n)
N=unique(n);
a=sum(n==N(1));
b=nchoosek(N(2:end),a);
try
c=b(randi(length(b)),:);
catch
flag=0;
break;
end
s=[s,n(n==N(1))];
t=[t,c];
n=n(a+1:end);
for k=1:a
try
n(find(n==c(k),1))=[];
catch
flag=0;
break;
end
end
end
if flag
g=graph(s,t);
plot(g);
m=full(adjacency(g));
break;
end
end
NeGM
NeGM le 13 Mai 2021
Hey David, Thank you it works pretty well!

Connectez-vous pour commenter.

Plus de réponses (1)

Matt J
Matt J le 13 Mai 2021
Modifié(e) : Matt J le 13 Mai 2021
Perhaps as follows,
n=6;
T=nan(n,2);
map=isnan(T);
while any(map(:))
[i,~]=find(map,1);
if ~map(i,1)
exclude=[i,T(i,1)];
elseif ~map(i,2)
exclude=[i,T(i,2)];
else
exclude=i;
end
pool=setdiff(find(any(map,2)),exclude);
if isempty(pool)
T=nan(n,2); map=isnan(T); continue
end
p=numel(pool);
j=pool(randi(p));
if map(i,1),
T(i,1)=j;
map(i,1)=0;
else,
T(i,2)=j;
map(i,2)=0;
end
if map(j,1),
T(j,1)=i;
map(j,1)=0;
else,
T(j,2)=i;
map(j,2)=0;
end
end
s=[(1:n), (1:n)].';
t=[T(:,1); T(:,2)];
st=unique(sort([s,t],2),'rows');
G=graph(st(:,1),st(:,2));
degree(G)
ans = 6×1
2 2 2 2 2 2

3 commentaires

NeGM
NeGM le 13 Mai 2021
hi Matt,
Under your code each node has a degree of 4, is there a way to contraint the connection to only have a specified number? Thank you
Matt J
Matt J le 13 Mai 2021
Fixed it.
NeGM
NeGM le 13 Mai 2021
Thank you

Connectez-vous pour commenter.

Catégories

En savoir plus sur Deep Learning Toolbox dans Centre d'aide et File Exchange

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by