for loop returns NaN

Question

negin tebyani 2018-2-10

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/381866-for-loop-returns-nan

回答： Walter Roberson 2018-2-10

I am trying to do this in my code:

for i=1:c+1
    load(row) = load(row)+(ro(row,i)* MinRateCurrentUserRRH(row,i))/(log(1+SINR(row,i)));
end

ro is a matrix of 1 and 0 s and MinRateCurrentUserRRH and SINR s are other matrixes. before running the for loop, load (row) is fine and for example 0.3, after first loop, it turns to NaN, so it gives errors, I can't figure out the problem. can anyone help me with it?

3 个评论
显示 1更早的评论隐藏 1更早的评论

negin tebyani 2018-2-10

编辑：negin tebyani 2018-2-10

SINR and currnetminrate are zero matrices that are filled in the code in other loops, in the time of running this, they look like:

and

and c is 1.

dpb 2018-2-10

在 MATLAB Online 中打开

As other poster says, nothing anybody here can do without actual data but you've got all the information you need; just use the debugger and see where your logic fails.

While not the specific problem, you can probably replace all the loops with Matlab array operations; "the MATLAB way". Look at the "dot" operators under

>> help punct
Punctuation.
.   Decimal point. 325/100, 3.25 and .325e1 are all the same.
.   Array operations.  Element-by-element multiplicative operations
    are obtained using .* , .^ , ./ , .\ or .'.  For example,
    C = A ./ B is the matrix with elements c(i,j) = a(i,j)/b(i,j).
  ...

请先登录，再进行评论。

请先登录，再回答此问题。

Answer 1

Walter Roberson 2018-2-10

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/381866-for-loop-returns-nan#answer_304501

Your row is a vector, so (log(1+SINR(row,i)) is a vector, so your / operator is mrdivide, the matrix right division operator, with A/B being approximately equivalent to A * pinv(B) where the * indicates mtimes, the algebraic matrix multiplication operator. For vector row you need the ./ operator instead of /

You will also need to watch out for the * operator earlier in the line: you probably want it to be .*