How to determine number of edges ?

Hi, below is the code for shortestpath:

 %code
A = randi(1,8)  ;
D = graph(A);   
D.Edges.W3 = randi(13, numedges(D), 1); 
D.Edges.Weight = D.Edges.W3;
p = plot(D,'o','EdgeLabel',D.Edges.Weight  );
axis square;  
grid on 
path = shortestpath(D, 2, 5);
highlight(p,path,'NodeColor','g','EdgeColor','g')
  end

The Result:

As the result shown that each of the point will generate 8 edges based on the "A = randi(1,8)" input. How can i determine the edges number manually?

Thanks.

[Walter Roberson]

Are you aware that randi(1,8) means the same as randi([1 1], 8, 8) which creates an 8 by 8 matrix in which each entry is a random integer in the range 1 to 1? So that line is the same as A = ones(8,8)

[Asyran Abdullah]

Ok, so it cant be configured manually right?

[Walter Roberson]

When you call graph() with a single argument then MATLAB assumes that you are passing in an adjacency matrix. The all-one's adjacency matrix connects every node to every node.

You can use

A = randi([0 1], 8, 8);
A = or(A, A.');
A = A - diag(diag(A));

to construct a random symmetric adjacency matrix with no self connections. You might find it a bit denser than you prefer, so there are other ways to generate it.

[Asyran Abdullah]

Ok, im edit some command and do this,

numNodes  =  12;
numEdge = 16;          
p = rand(numNodes,2);
x = p(:,1)*100;
y = p(:,2)*100;
st = unique( sort( randi(numNodes, numEdge, 2), 2), 'rows');

[Walter Roberson]

Yes, looks like the code I suggested in https://www.mathworks.com/matlabcentral/answers/416045-how-to-make-a-line-between-random-nodes-connection#answer_333928

But I am not certain what your question is? Are you trying to control the number of total edges that are created? Are you trying to control the minimum "degree" of each node? Are you trying to control the maximum degree? Are you trying to determine the degree of each node on the path between nodes 2 and 5 ?

[Asyran Abdullah]

yes, based on your code. i try to determine the degree of each node, but nevermind i already get it. Thanks sir.

Answer 1

[Image Analyst]

Since every node seems to go to every node (even including itself), wouldn't it just be simply N*N/2 = 8*8/2 = 32?

[Asyran Abdullah]

Is there any way to control or avoid duplicate edges?

[Image Analyst]

Which of those edges is duplicated?

[Asyran Abdullah]

I found this to prevent self loop

% 
G = rmedge(G, 1:numnodes(G), 1:numnodes(G));

[Walter Roberson]

The line I showed,

A = A - diag(diag(A));

prevents edges between a node and itself, provided that A is a square adjacency matrix.