Can you solve this in matlab2009b 64 bit? (systems of equations to be solved using "solve")
in [Matlab]
|
Prev: is there a way to automatically read in all the sheet names in an Excel book?
Next: Support Vector Machines
From: Edward D. on 22 Jan 2010 18:35 Hi all, Can you solve the following in matlab 2009b 64 bit (macosx with snow leopard)? I get the errors given below the code. It takes a lot of time before Matlab gives the error. Where do I go wrong? %----------------------- syms p0 p00 p1 p2 cn cr dn dr a an ar bn br T tau real; myfun = [ -(tau*(2*an + 2*bn*cn - 4*bn*p1 - dn*tau))/2; -(tau*(2*ar + 2*br*cr - 4*br*p0 - dr*tau))/2; ((T - tau)*(T*dn - 2*an - 2*bn*cn + 4*bn*p2 + dn*tau))/2; ((T - tau)*(T*dr - 2*ar - 2*br*cr + 4*br*p00 + dr*tau))/2; an*p2 - an*p1 - ar*p0 + ar*p00 + bn*p1^2 - bn*p2^2 + br*p0^2 - br*p00^2 - bn*cn*p1 + bn*cn*p2 - br*cr*p0 + br*cr*p00 + dn*p1*tau - dn*p2*tau + dr*p0*tau - dr*p00*tau]; soln = solve(myfun,'p1','p2','p0','p00','tau'); %----------------------- ??? Error using ==> mupadmex Error in MuPAD command: Index exceeds matrix dimensions. Error in ==> sym.sym>sym.subsref at 1366 B = mupadmex('mllib::subsref',A.s,inds{:}); Error in ==> solve>assignOutputs at 110 S.(char(symvars(j))) = R(:,j); Error in ==> solve at 87 varargout = assignOutputs(nargout,R,symvars);
From: Bruce on 30 Jan 2010 21:23
"Edward D." <edwardawsremove.this(a)gmail.com> wrote in message <hjdcnp$55c$1(a)fred.mathworks.com>... > Hi all, > > Can you solve the following in matlab 2009b 64 bit (macosx with snow leopard)? I get the errors given below the code. It takes a lot of time before Matlab gives the error. Where do I go wrong? > > %----------------------- > syms p0 p00 p1 p2 cn cr dn dr a an ar bn br T tau real; > > myfun = [ -(tau*(2*an + 2*bn*cn - 4*bn*p1 - dn*tau))/2; -(tau*(2*ar + 2*br*cr - 4*br*p0 - dr*tau))/2; ((T - tau)*(T*dn - 2*an - 2*bn*cn + 4*bn*p2 + dn*tau))/2; ((T - tau)*(T*dr - 2*ar - 2*br*cr + 4*br*p00 + dr*tau))/2; an*p2 - an*p1 - ar*p0 + ar*p00 + bn*p1^2 - bn*p2^2 + br*p0^2 - br*p00^2 - bn*cn*p1 + bn*cn*p2 - br*cr*p0 + br*cr*p00 + dn*p1*tau - dn*p2*tau + dr*p0*tau - dr*p00*tau]; > > soln = solve(myfun,'p1','p2','p0','p00','tau'); > %----------------------- > > ??? Error using ==> mupadmex > Error in MuPAD command: Index exceeds matrix dimensions. > > Error in ==> sym.sym>sym.subsref at 1366 > B = mupadmex('mllib::subsref',A.s,inds{:}); > > Error in ==> solve>assignOutputs at 110 > S.(char(symvars(j))) = R(:,j); > > Error in ==> solve at 87 > varargout = assignOutputs(nargout,R,symvars); Hello Edward, I used Matlab r2009B to try out the problem on a 32-bit computer. I could not get the very last bit of the matrix to work but the rest worked fine. You just need to enclose the expressions in single quotes. I converted the solution from a struct to a vector and then displayed the solution as below. clc; syms p0 p00 p1 p2 cn cr dn dr a an ar bn br T tau real; myfun = [ '-(tau*(2*an + 2*bn*cn - 4*bn*p1 - dn*tau))/2', ... '-(tau*(2*ar + 2*br*cr - 4*br*p0 - dr*tau))/2', ... '((T - tau)*(T*dn - 2*an - 2*bn*cn + 4*bn*p2 + dn*tau))/2', ... '((T - tau)*(T*dr - 2*ar - 2*br*cr + 4*br*p00 + dr*tau))/2']; %, ...last bit is an invalid expression %'an*p2 - an*p1 - ar*p0 + ar*p00 + bn*p1^2 - bn*p2^2 + br*p0^2 - br*p00^2... %- bn*cn*p1 + bn*cn*p2 - br*cr*p0 + br*cr*p00 + dn*p1*tau - dn*p2*tau + dr*p0*tau - dr*p00*tau']; soln = solve(myfun,'p1','p2','p0','p00','tau'); %turn structure into row vector then take transpose for easy reading soln = [soln.p1 soln.p2 soln.p0 soln.p00 soln.tau]'; pretty(soln); +- -+ | conj(x), conj(u) | | | | conj(u), conj(v) | | | | 8 an + 2 ar + 8 bn cn + 2 br cr - 16 bn conj(x) - 4 dn conj(v) - dr conj(v) | | ---------------------------------------------------------------------------, conj(y) | | 4 br | | | | conj(y), conj(x) | | | | conj(v), 0 | +- -+ Published with MATLAB® 7.9 ============================================= Regards Bruce |