Prev: For loop problem
Next: Vectorize AR Stochastic Process
From: Richard on 13 Aug 2010 08:30
Hi I have a system of linear equations(that I've finally vectorized.).I need to solve the following for P but I'm lost as to how to extract P from the vectorized formula to solve. I've put it on 3 lines for readability. Not even sure if I'm providing enough info here?
Any help appreciated!
Thanks
SIRtarget and eta are scalar values
U is an n*1 column
G and X are n*m matrices
P is a row of length m

SIRtarget.*(repmat(U,1,length(G(1,:))).*G)'*(sum(repmat(P,length(X(:,1)),1).*(X./G),2))-SIRtarget.*P'
+SIRtarget*eta*ones(length(G(1,:)),1)
-P'

here is the same formula in loop form so its more readable:

for j=1:length(A)
msum=0;
for h=1:n
s=0;
for t=1:m
s=s+(P(t)*X(h,t)/G(h,t));
end
msum=msum+U(h)*G(h,j)*s;
end
A(j)=msum;
end
SIRtarget.*A - SIRtarget.*P' + SIRtarget.*eta - P'
From: Richard on 13 Aug 2010 10:44
"Richard " <REMOVETHISrcaldwellie(a)yahoo.com> wrote in message <i43dss$863$1(a)fred.mathworks.com>...
> Hi I have a system of linear equations(that I've finally vectorized.).I need to solve the following for P but I'm lost as to how to extract P from the vectorized formula to solve. I've put it on 3 lines for readability. Not even sure if I'm providing enough info here?
> Any help appreciated!
> Thanks
> SIRtarget and eta are scalar values
> U is an n*1 column
> G and X are n*m matrices
> P is a row of length m
>
> SIRtarget.*(repmat(U,1,length(G(1,:))).*G)'*(sum(repmat(P,length(X(:,1)),1).*(X./G),2))-SIRtarget.*P'
> +SIRtarget*eta*ones(length(G(1,:)),1)
> -P'
>
> here is the same formula in loop form so its more readable:
>
> for j=1:length(A)
> msum=0;
> for h=1:n
> s=0;
> for t=1:m
> s=s+(P(t)*X(h,t)/G(h,t));
> end
> msum=msum+U(h)*G(h,j)*s;
> end
> A(j)=msum;
> end
> SIRtarget.*A - SIRtarget.*P' + SIRtarget.*eta - P'

No answers? Am I trying to do something impossible or is my post just too hard to read?
From: Greg Heath on 13 Aug 2010 20:27
On Aug 13, 10:44 am, "Richard " <REMOVETHISrcaldwel...(a)yahoo.com>
wrote:
> "Richard " <REMOVETHISrcaldwel...(a)yahoo.com> wrote in message <i43dss$86....(a)fred.mathworks.com>...
> > Hi I have a system of linear equations(that I've finally vectorized.).I  need to solve the following for P but I'm lost as to how to extract P from the vectorized formula to solve. I've put it on 3 lines for readability.. Not even sure if I'm providing enough info here?
> > Any help appreciated!
> > Thanks
> > SIRtarget and eta are scalar values
> > U is an n*1 column
> > G and X are n*m matrices
> > P is a row of length m
>
> > SIRtarget.*(repmat(U,1,length(G(1,:))).*G)'*(sum(repmat(P,length(X(:,1)),1)­.*(X./G),2))-SIRtarget.*P'
> > +SIRtarget*eta*ones(length(G(1,:)),1)
> > -P'
>
> > here is the same formula in loop form so its more readable:
>
> > for j=1:length(A)
> >      msum=0;
> >      for h=1:n
> >          s=0;
> >          for t=1:m
> >              s=s+(P(t)*X(h,t)/G(h,t));
> >          end
> >          msum=msum+U(h)*G(h,j)*s;
> >      end
> >      A(j)=msum;
> >  end
> >  SIRtarget.*A - SIRtarget.*P' + SIRtarget.*eta - P'
>
> No answers? Am I trying to do something impossible or is my post just too hard to read

Where is your system linear equations? I am looking for something that
looks like

A*x = b

Hope this helps.

Greg
 | 
Pages: 1
Prev: For loop problem
Next: Vectorize AR Stochastic Process