Product Suggestions

Post your suggestions on new features and products.

The XMLTools package appears to use a DOM parser.  This requires that all of the XML data be loaded into memory and parsed before it can be used.  This is less efficient than a SAX or stream-based parser that operates on only parts of the XML tree at a time.  It would be a useful enhancment to include SAX based parsing in Maple.

Hi, have a look at the pic. Maple has occasionally problems to combine solutions of an integral (and not only there) to a real term. One has to  do it by hand.  Are there any plans to improve this behaviour ?

 

Hi,

I just realized that Maple 8 shuts down without saving every time I try to save a document. Obviously, this is a very serious problem. Any tips? I actually uninstalled the whole software package, and reinstalled it, only to find that the problem was still there. I hope someone else has dealt with this. Please help. Thanks,

John

Bonjour,

It seems that bit to bit functions on integers, like the C-functions and, or, not, xor, << (left shift) , >> (right shift), are missing in Maple.
 

Why?

> 8 and 25;
Error, invalid boolean expression

 

GraphTheory includes ImportGraph and ExportGraph.  These are handy functions but they would be much more useful if they supported GraphML.  GraphML is a popular standard.  It is particularly useful for integration with tools built with the Prefuse API (http://prefuse.org/).  Please add GraphML support for importing and exporting Graphs.

As Demmel and others have noted, SVD is both more reliable and more expensive than QR as a method of solving rank-deficient least squares problems.

SVD is the method that LinearAlgebra:-LeastSquares will choose when the Matrix has more columns than rows (n>m), unless instructed otherwise using the optional 'method' parameter.

LinearAlgebra:-SingularValues always computes a full U and Vt. But for least squares computations, such as when n>m, this is not necessary. Including the smaller singular values may just be (re-)introducing noise. See here for more detail.

Here's a 20x2000 example, using wrapperless external calling and the SVD routine dgesvd in the CLAPACK library. The effective speedup by using the Thin SVD for that 20x2000 least squares example is about a factor of 100 (ie, 2000/20), with a similar reduction in additional memory allocation.

You have options to open tabs in separate windows or move existing multiple tabs into separate windows. 

Can we add an option to combine tabs into one window? 

  restart: interface(version); Digits:=14:

    Classic Worksheet Interface, Maple 11.02, Windows, Nov 10 2007 Build ID 330022

  Ei(1,1/2*Pi*(1+2*k)): 
  %=convert(%,Sum);
  subs(k=0,%);
  evalf(%);

                    Pi (1 + 2 k)
              Ei(1, ------------) = GAMMA(Pi k + 1/2 Pi)
                         2

                             Pi            Pi
                      Ei(1, ----) = GAMMA(----)
                             2            ...

I have been wondering about the reception of an independent patch Library for Maple.

I'll lay out a few ideas, and then maybe some criticism (or indifference) might follow.

I'm a bullet-point sort of person:

  • sourceforge project with a few willing experts for vetting and gatekeeping of submissions.
  • Patch .mla Library archive, going into /toolbox/Patch/lib/ so as to get picked up automatically by libname.
  • Self-building, with...

Does anyone here use this?

Would it be useful for someone new to topology/homological algebra learning wise ?

comments/opinions very welcomed.

thanks,

dc

http://wwwb.math.rwth-aachen.de/homalg/

There have been a few posts on mapleprimes about numerically solving systems of procedures. The latest one, up until now, was this.

Here's some code to implement the method. Since the algorithm is basically very simple, I've added a few bells and whistles as optional arguments.

The essence of it is as follows. The number of procedures must match the number of parameters of each and every procedure. It does maxtries attempts at choosing a random point, and then does at most maxiter iterations. A solution is only accepted if the norm of the last change in vector (point) x is less than xtol, and if the forward error norm(F(x)) is less than ftol. The jacobian of F may be supplied optionally as a Matrix of procedures, or a method for computing the jacobian may be supplied. The methods are fdiff which only uses Maple's numerical differentiation routine fdiff, or hybrid which attempts symbolic differentiation via Maple's D[] operator and then falls back to fdiff via the nifty evalf@D equivalence.

I have not profiled it, though I have attempted to set up the jacobian construction so that it re-uses the same rtable for each instantiation. I have not gone really low-level with the external-calling to numeric solvers. I have not tried to leverage fast hardware double-precision solutions as jumping-off points for extended precision polishing (eg. see here).

I've added some examples, after the code below. Feel free to comment or mention bugs. I'd really like a better name for it.

 

 

restart:

NRprocs:=proc(

  funlist::{list(procedure),Vector(procedure)}

, {allowcomplex::truefalse:=false}

, {initialpoint::Vector:=NoUserValue}

, {maxtries::posint:=NoUserValue}

, {maxiter::posint:=30}

, {xtol::realcons:=5.0*10^(-Digits)}

, {ftol::realcons:=1.0*10^(-trunc(2*Digits/3))}

, {method::{identical(hybrid,fdiff)}:=':-hybrid'}

, {jacobian::{Matrix({operator,procedure}),identical(NoUserValue)}:=':-NoUserValue'}

, $

)

local a, b, currXseq, dummy, dummyseq, F, i, ii, J, j, jj,

      MaxTries, numvars, oldX, thisJ, thistry, varnumset, X;

 

  varnumset := {seq(`if`(not(member('call_external',{op(3,eval(F))})),
                         nops([op(1,eval(F))]),NULL), F in funlist)};

  if nops(varnumset) > 1 then

    error "expecting procedure all taking same number of arguments";

  elif nops(varnumset) = 1 then

    numvars := varnumset[1];

  else

    numvars := nops(funlist);

  end if;

 

  if numvars <> nops(funlist) then

    error "number of procedures %1 does not match number of their parameters %2"

, nops(funlist), numvars;

  end if;

 

  if initialpoint <> 'NoUserValue' then

    oldX := Vector(numvars,(i)->evalf(initialpoint[i]),':-storage'=':-rectangular',

                   ':-datatype'=`if`(allowcomplex,'complex'(':-float'),':-float'));

    if maxtries = ':-NoUserValue' then

      MaxTries := 1;

    elif maxtries > 1 then

      error "maxtries must be 1 when initialpoint is supplied";

    else

      MaxTries := maxtries;

    end if;

  else

    if maxtries = ':-NoUserValue' then

      MaxTries := 20;

    else

      MaxTries := maxtries;

    end if;

  end if;

 

  if jacobian = ':-NoUserValue' then

    # This is only built once, so there should be a way to make

    # it take a little more time and make a J that evaluates quicker.

    # Using evalhf is just one possibility.

    dummyseq:=seq(dummy[i],i=1..numvars);

    if method=':-hybrid' then

      J:=Matrix(nops(funlist),numvars,

              (i,j)->unapply('evalf'(D[j](funlist[i])(dummyseq)),[dummyseq]));

    else # method=fdiff

      J:=Matrix(nops(funlist),numvars,

              (i,j)->unapply(fdiff(funlist[i],[j],[dummyseq]),[dummyseq]));

    end if;

  else

    J:=jacobian;

  end if;

  thisJ := Matrix(nops(funlist),numvars,

                  ':-datatype'=`if`(allowcomplex,'complex'(':-float'),':-float'));

                                                                                

  X := Vector(numvars,':-datatype'=`if`(allowcomplex,'complex'(':-float'),':-float'));

 

  userinfo(1,'NRprocs',`maximal number of tries is `, MaxTries);

  for thistry from 1 to MaxTries do

 

    if thistry > 1 or initialpoint = ':-NoUserValue' then

      if allowcomplex = true then

        oldX := LinearAlgebra:-RandomVector(numvars,

                              ':-density'=1,':-generator'=-1.0-1.0*I..1.0+1.0*I,

                              ':-outputoptions'=[':-datatype'='complex'(':-float')]);

      else

        oldX := LinearAlgebra:-RandomVector(numvars,

                              ':-density'=1,':-generator'=-1.0..1.0,

                              ':-outputoptions'=[':-datatype'=':-float']);

      end if;

    end if;

 

    userinfo(1,'NRprocs',`initial point : `, convert(oldX,list));

    userinfo(1,'NRprocs',`maximal number of iterations is `, maxiter);

    try

      for ii from 1 to maxiter do

        currXseq:=seq(oldX[jj],jj=1..numvars);

        for i from 1 to nops(funlist) do

          for j from 1 to numvars do

            thisJ[i,j] := J[i,j](currXseq);

          end do;

        end do;

        # copy oldX into X, so that increment can be added to X inplace.

        ArrayTools:-Copy(numvars,oldX,0,1,X,0,1);

        LinearAlgebra:-LA_Main:-VectorAdd(

               X, LinearAlgebra:-LA_Main:-LinearSolve(

                   thisJ,

                   Vector(nops(funlist),

                     (i)->evalf(funlist[i](currXseq))),

                   ':-method'=':-none',':-free'=':-NoUserValue',':-conjugate'=true,

                   ':-inplace'=false,':-outputoptions'=[],':-methodoptions'=[]),

               1,-1,':-inplace'=true,':-outputoptions'=[]);

        userinfo(2,'NRprocs',`at iteration`,ii, convert(X,list));

        if LinearAlgebra:-LA_Main:-Norm(

             # oldX is no longer needed, so can overwrite it inplace.

             LinearAlgebra:-LA_Main:-VectorAdd(

               oldX,X,-1,1,':-inplace'=true,':-outputoptions'=[]),

             infinity,':-conjugate'=true) < xtol

         and max(seq(abs(F(seq(X[jj],jj=1..numvars))),F in funlist)) < ftol then

          return X;

        end if;

        ArrayTools:-Copy(numvars,X,0,1,oldX,0,1);

      end do:

    catch "singular matrix","inconsistent system","unable to store":

      ii := ii - 1;

      next;

    end try;

  end do;

  return 'procname'(args);

end proc:

f:=proc(x,y) x^3-sin(y); end proc:

g:=proc(x,y)

local sol;

  sol:=fsolve(q^5+x*q^2+y*q=0,q,complex):

  Re(sol[1]+1);

end proc:

                                                                                

fl:=[f,g]:

NRprocs(fl);

Vector[column]([[-.897503424482696], [3.94965547519010]])

if type(%,Vector) then

  max(seq(abs(fl[i](seq(%[j],j=1..nops(fl)))),i=1..nops(fl)));

end if;

HFloat(8.881784197001252e-16)

 

Digits:=32:

NRprocs(fl);

if type(%,Vector) then

  max(seq(abs(fl[i](seq(%[j],j=1..nops(fl)))),i=1..nops(fl)));

end if;

Digits:=10:

Vector[column]([[-.89750342448269837150872333768519], [3.9496554751901143338368997374538]])

0.1e-31

 

infolevel[NRprocs]:=1:

sol := NRprocs([f,g],initialpoint=<5.0, 6.0>);

if type(%,Vector) then

  max(seq(abs(fl[i](seq(%[j],j=1..nops(fl)))),i=1..nops(fl)));

end if;

infolevel[NRprocs]:=0:

NRprocs: maximal number of tries is  1
NRprocs: initial point :  [HFloat(5.0) HFloat(6.0)]
NRprocs: maximal number of iterations is  30

sol := Vector(2, {(1) = -.8975034244826993, (2) = 3.9496554751901174}, datatype = float[8])

HFloat(4.440892098500626e-16)

 

infolevel[NRprocs]:=2:

sol := NRprocs(fl,initialpoint=<-1, -2>);

if type(%,Vector) then

  max(seq(abs(fl[i](seq(%[j],j=1..nops(fl)))),i=1..nops(fl)));

end if;

infolevel[NRprocs]:=0:

NRprocs: maximal number of tries is  1

NRprocs: initial point :  [HFloat(-1.0) HFloat(-2.0)]
NRprocs: maximal number of iterations is  30
NRprocs: at iteration 1 [HFloat(-0.973448865779436) HFloat(-1.973448865779436)]
NRprocs: at iteration 2 [HFloat(-0.9727013515968866) HFloat(-1.9727013515968865)]
NRprocs: at iteration 3 [HFloat(-0.9727007668592758) HFloat(-1.9727007668592722)]
NRprocs: at iteration 4 [HFloat(-0.9727007668589176) HFloat(-1.972700766858918)]

sol := Vector(2, {(1) = -.9727007668589176, (2) = -1.972700766858918}, datatype = float[8])

HFloat(1.1102230246251565e-16)

sol := NRprocs(fl);

if type(%,Vector) then

  max(seq(abs(fl[i](seq(%[j],j=1..nops(fl)))),i=1..nops(fl)));

end if;

sol := Vector(2, {(1) = -.9727007668589175, (2) = -1.9727007668589187}, datatype = float[8])

HFloat(6.661338147750939e-16)

sol := NRprocs(fl,ftol=1e-8);

if type(%,Vector) then

  max(seq(abs(fl[i](seq(%[j],j=1..nops(fl)))),i=1..nops(fl)));

end if;

sol := Vector(2, {(1) = -.8975034244826969, (2) = 3.9496554751901094}, datatype = float[8])

HFloat(3.3306690738754696e-16)

sol := NRprocs(fl,initialpoint=<-I,I>,allowcomplex=true);

if type(%,Vector) then

  max(seq(abs(fl[i](seq(%[j],j=1..nops(fl)))),i=1..nops(fl)));

end if;

sol := Vector(2, {(1) = -.13507476585608832-.9226174570773964*I, (2) = 2.865935679636024-.7040591486912059*I}, datatype = complex[8])

HFloat(2.268185639309195e-12)

sol := NRprocs(fl,allowcomplex=true);

if type(%,Vector) then

  max(seq(abs(fl[i](seq(%[j],j=1..nops(fl)))),i=1..nops(fl)));

end if;

sol := Vector(2, {(1) = .3864813801943507-.6844281432614738*I, (2) = -.5067551915310455+0.15920328832212956e-1*I}, datatype = complex[8])

HFloat(8.281264562981505e-12)

# deliberate errors
NRprocs([f,g,(a,b)->a+b]);

NRprocs([(a,b,c)->a*b*c]);

Error, (in NRprocs) number of procedures 3 does not match number of their parameters 2

Error, (in NRprocs) number of procedures 1 does not match number of their parameters 3

fl:=[x->cos(x)-x]:

CodeTools:-Usage( NRprocs(fl) );

if type(%,Vector) then

  max(seq(abs(fl[i](seq(%[j],j=1..nops(fl)))),i=1..nops(fl)));

end if;

memory used=372.34KiB, alloc change=0 bytes, cpu time=15.00ms, real time=7.00ms, gc time=0ns

Vector[column]([[.739085133215161]])

HFloat(0.0)

fl:=[(x,y,z)->x*z+2*y^2-sqrt(z),(x,y,z)->z+x*y,(x,y,z)->x^3+y*z]:

CodeTools:-Usage( NRprocs(fl) );

if type(%,Vector) then

  max(seq(abs(fl[i](seq(%[j],j=1..nops(fl)))),i=1..nops(fl)));

end if;

memory used=3.28MiB, alloc change=0 bytes, cpu time=63.00ms, real time=66.00ms, gc time=0ns

Vector[column]([[.414213562373095], [-.414213562373095], [.171572875253810]])

HFloat(1.1102230246251565e-16)

jfl:=Matrix(3,3,(i,j)->D[j](fl[i])):

CodeTools:-Usage( NRprocs(fl,jacobian=jfl) );

if type(%,Vector) then

  max(seq(abs(fl[i](seq(%[j],j=1..nops(fl)))),i=1..nops(fl)));

end if;

memory used=0.64MiB, alloc change=0 bytes, cpu time=15.00ms, real time=13.00ms, gc time=0ns

Vector[column]([[.414213562373095], [-.414213562373095], [.171572875253810]])

HFloat(5.551115123125783e-17)

fl:=[proc(x::float,y::float,z::float) x*z+1.9*y^2-z^2; end proc,

     proc(x::float,y::float,z::float) z+0.9*x*y; end proc,

     proc(x::float,y::float,z::float) x^3+0.9*y*z; end proc]:

st,ba,bu:=time(),kernelopts(bytesalloc),kernelopts(bytesused):

jfl:=Matrix(3,3,(i,j)->D[j](fl[i])):

sol:=CodeTools:-Usage( NRprocs(fl,jacobian=jfl,maxtries=50) );

if type(sol,Vector) then

  max(seq(abs(fl[i](seq(sol[j],j=1..nops(fl)))),i=1..nops(fl)));

end if;

memory used=9.20MiB, alloc change=0 bytes, cpu time=187.00ms, real time=182.00ms, gc time=0ns

sol := Vector(3, {(1) = 2.1111111111111107, (2) = -2.3456790123456783, (3) = 4.4567901234567895}, datatype = float[8])

HFloat(3.552713678800501e-15)

fl:=[proc(x::float,y::float,z::float) x*z+1.8*y^2-z^2; end proc,

     proc(x::float,y::float,z::float) z+0.8*x*y; end proc,

     proc(x::float,y::float,z::float) x^3+0.8*y*z; end proc]:

jfl:=Matrix(3,3,(i,j)->D[j](fl[i])):

cfl:=[seq(Compiler:-Compile(fl[i]),i=1..nops(fl))]:

cjfl:=map(t->`if`(type(t,procedure),Compiler:-Compile(t),t),jfl):

sol:=CodeTools:-Usage( NRprocs(cfl,jacobian=cjfl) );

if type(sol,Vector) then

  max(seq(abs(fl[i](seq(sol[j],j=1..nops(fl)))),i=1..nops(fl)));

end if;

memory used=41.43MiB, alloc change=0 bytes, cpu time=920.00ms, real time=884.00ms, gc time=78.00ms

sol := Vector(3, {(1) = -1.2499999999999998, (2) = 1.5624999999999998, (3) = 1.5624999999999996}, datatype = float[8])

HFloat(4.440892098500626e-16)

NRprocs([f,g],method=hybrid);

Vector[column]([[-.972700766858918], [-1.97270076685891]])

[f, g](seq(a, a=%)); # check whether it's a root

[HFloat(0.0), HFloat(7.771561172376096e-16)]

NRprocs([f,g],method=fdiff);

Vector[column]([[-.972700766858918], [-1.97270076685892]])

[f, g](seq(a, a=%)); # check whether it's a root

[HFloat(-1.1102230246251565e-16), HFloat(0.0)]

 

 

Download fsolve_procs.mw

acer

BUG 1:

Severity: Medium-High

Reproduction #1:
>with(Student([Calculus1]));

IntTutor;

Try to enter "csc(x)^3" in "function", and let variable to be "x"

Click all steps, and watch the system freezes.
Or yo can click "Next Step", which won't change the integrand at all, you may click 100000000 times -- the program is in an infinite loop.

If one has ever used an XML editor, one knows the frustration of you having to do the bulk of the work locating mismatched or unclosed <> tags. Let us view tags as delimiters on sets and subsets. Consider the case where the XML editor has checked all beginning and ending tags and found an equal number of beginning and endint tags of a given name: a, b, etc. Nevertheless, even though we have an equal number of and tags and and tags is wrong syntactically because the "set" "intersects" with the "set" non-trivially, yet neither set is a proper subset of the other. Certainly, if a computer algebra system such as Maple can determine set overlaps like this, then
MapleTA currently restricts feedback in only one sense -- you can prevent access to feedback until AFTER a specified date. I provide MapleTA support for the University of Nebraska-Lincoln, and an instructor has come to me with what seems like a reasonable desire -- that unfortunately does not seem possible at this point. He is creating proctored exams. Students take proctored exams on campus in our testing center, which is staffed by proctors. The instructor wants his students to view the results of their exams in this proctored setting so that they can mentally take note of which questions they got wrong and which questions they got right, but cannot produce a hard copy of the exam. And he doesn't want students to be able to go back to their dorm rooms and view the exam.
I was just wondering what a good book to learn engineering related mathematics with maple would be? There's one on the website of maplesoft but 125 seems a lot to pay for a book I'll use just for fun. So do you guys have any suggestions? Thnaks
First 15 16 17 18 19 20 21 Page 17 of 23