A quick way to make sure the columns in each row are decreasing for matrix A

Hello everyone,

I have a matrix A with 100 rows and 20 columns.

I would like to ensure that for evry and each row "i", the value in column "j" is not greater than the value in row i and column "j-1". If it is, then replace it with the corresponding value in row i, column j-1 minus a very small value 0.001.

For example, if

A= [ 4,  2.1, 2.2101, 1;
    20, 10,   5,      1]

. Here, for simplicity, I only consider A to have 2 rows and 4 columns.

In this example we see that for row 1, the value in column 3 is 2.2101 which is greater than the value in row 1, column 2 that is 2.1. So, I will modify A(1, 3)=A(1,2)-0.001= 2.099. Therefore, A_new= [4 2.1 2.0.099 1; 20 10 5 1].

Answer 1

[dpb]

That'll take iterating over the columns...and could be necessary to be recursive depending upon the input data. But, for the one case of the example;

A= [ 4,  2.1, 2.2101, 1;
    20, 10,   5,      1];
for i=2:size(A,2)
  ix=(A(:,i)>A(:,i-1));
  if any(ix), A(ix,i)=A(ix,i-1)-0.001; end
end
disp(A)

    4.0000    2.1000    2.0990    1.0000
   20.0000   10.0000    5.0000    1.0000

[H R]

Thank you