On Wednesday, 2 July 2014 19:36:51 UTC+1, Axel Vogt  wrote:   
   > On 02.07.2014 19:35, Axel Vogt wrote:   
   >   
   > > On 02.07.2014 14:34, paulandrewbird@gmail.com wrote:   
   >   
   > >> On Tuesday, 1 July 2014 20:43:02 UTC+1, Axel Vogt  wrote:   
   >   
   > >>> Do you have any (numerical) examples where your guess might be correct?   
   >   
   > >>   
   >   
   > >> My guess is simply based on the similarity to the quadratic integral:   
   >   
   > >>   
   >   
   > >> Integral[ exp(-A_ij x_i x_j )]dx^n = sqrt(pi^n / det(A) )   
   >   
   > >   
   >   
   > > If you mean the sum: for matrix = unit matrix then it is   
   >   
   > > exp( x^2 + y^2 ) to be integrated over the plane, which   
   >   
   > > is infinity, while det = 1. So at least you need sign   
   >   
   > > conditions ( A = - Id will do)   
   >   
   >   
   >   
   > PS: there is some issue with your formula   
   >   
   >   
   >   
   >          [-1     1]   
   >   
   > for A = [        ]  having det = +1 that means to integrate   
   >   
   >          [ 0    -1]   
   >   
   >   
   >   
   > exp( -x^2+x*y-y^2 ), resulting in Pi * 2/sqrt(3)   
      
   Yeah you'd have to assume a symmetric matrix or you can replace det(A) with   
   det( (A+A^T)/2 ).   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)