Generating a "special" probability random matrix in efficient way

Hi,

I would to generate a probability random matrix with two constraints: the first is that the sum of the values of the row must be equal to 1 and the second is that one element of each row has to be a value greater than 0.5. I have written this:

if true
  for l=1:n
  while (nnz(random_data(l,:)) < d)
      m=randi(d,1,1);
      if (random_data(l,m)==0)
          row_sum=sum(random_data(l,:));
          if(row_sum == 0)
              random_data(l,m)=0.5+0.5*rand(1,1);
          else
              if(nnz(random_data(l,:))==(d-1))
                  random_data(l,m)=1-row_sum;
              else
                  random_data(l,m)=(1-row_sum)*rand(1,1);
              end
          end
      end
  end
end
end
It is very inefficient. Can you help me?

Answer 1

try this

n=10,m=12;
x=rand(n,m)/m;
[~,id1]=max(x');
id2=m+1-id1;
for k=1:n
  x(k,id1(k))=0.5;
  x(k,id2(k))=1-sum(x(k,:))+x(k,id2(k))
end