How to discover the neighbor nodes in an matrix?

Hello guys,

I'm new here and I need your help. I have this code that I need to translate it into Matlab. I have an idea to create a matrix of distances between nodes, where for example: d11is the distanace between 1 and 1 (which will be 0), d12 is the distance of 1 with 2. Then checking the first row of the matrix compares the distance values with the transmission radius of node 1 and if it is smaller than the radius then these nodes will be included in the vector of neighbors of the first node. this is my idea but I don't know how to write it in Matlab. If someone could help me, I would be very grateful.

[John D'Errico]

I had to laugh when I read what you posted. Clearly, the authors do not test their own code. I see that sometimes they write declare, but at other times they use the declear tool. I wonder if there is a difference? I wonder what declear does anyway?

And is that Preducre, or Preduce? Or just maybe is it supposed to be Procedure? The point is, I doubt I would trust such a poorly written/edited/proofread text to be accurate.

[Jan]

John hits the point.

There is a "Neighbor list" as input, a "Neighbor_list" and empty array, a "MyNeighborsSet" and "Neighbors". The pseudocode is not useful and you cannot implement it in Matlab reliably.

Answer 1

[Jon]

Here is a little example that I think does what you are asking

%
% assign threshold for being a neighbor
dMax = 0.6;

% make an example distance matrix where i,j entry gives distance from node
% i to node j
D = rand(5,5);
% make the example matrix symmetric
D = (D + D')/2 
% make main diagonal zero, nodes are always neighbors of themselves
D = D - diag(diag(D)); 

% find the neighbors of each node
% form logical matrix whose (i,j)th entries are true if node j is a
% neighbor of node i
isNeighbor = D <= dMax; 

% make neighbor lists
% use cell array because lists may have different lengths
numNodes = size(D,1);
Neighbor = cell(numNodes,1);
for k = 1:numNodes
    Neighbor{k} = find(isNeighbor(k,:));
end
display(Neighbor)

You could also do this using MATLAB's graph functions, replacing the last lines of the above code with

% could also do this using a graph
G = graph(isNeighbor);
Neighbor = cell(numNodes,1);
for k = 1:numNodes
    Neighbor{k} = neighbors(G,k)';
end
display(Neighbor);

[Jon]

Note also that MATLAB graphs even allow finding nearest neighbors within a radius, so you could probably even further simplify the above.