On 24.09.13 14:05, El wrote: > Hello, > > I am having a problem using solve to find x that fulfil f(x)=0. My function is: > f=trace(A1*P)-SNR_linear(m); > temp=solve(f==0,x); > where A1 is given matrix which elements are known, P is diagonal matrix with diagonal elements containing x and I have to find x for m=1:length(SNR_linear). I am doing this for ten different matrix A1. > But if I use the obtained values of x to check the value of trace(A1*P) sometimes it is bigger than SNR_linear(m). Is it possible for these to be pathological cases?
Could you show a specific example, as short as possible?