[Maxima-bugs] [ maxima-Bugs-3558096 ] to_poly_solve gives a wrong solution for cos(x)=sin(3x)

[SourceForge.net <no...@so...>]

Bugs item #3558096, was opened at 2012-08-15 17:46
Message generated for change (Comment added) made by jyoberle
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=104933&aid=3558096&group_id=4933

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: Lisp Core - Solving equations
Group: None
Status: Open
Resolution: None
Priority: 5
Private: No
Submitted By: Jean-Yves (jyoberle)
Assigned to: Nobody/Anonymous (nobody)
Summary: to_poly_solve gives a wrong solution for cos(x)=sin(3x)

Initial Comment:
Hi,

When doing:
  load(to_poly_solve);
  algexact:true;
  to_poly_solve(cos(x)-sin(3*x),x);

I get:
  to_poly_solve: to_poly_solver.mac is obsolete; I'm loading to_poly_solve.mac instead.
  %union([x=-(4*%pi*%z0+%pi)/4],[x=(4*%pi*%z1+%pi)/8])

But I think that the first solution should be (based on hand solving): (4*%pi*%z0+%pi)/4 (no minus sign).

For example: if we consider %z0 = 0 in the to_poly_solve solution, we get x=-%pi/4 which is not a solution of the equation cos(x)-sin(3*x). On the other hand, if we set %z0 = 0 in the hand found solution, we get x=%pi/4 which is a solution.

The build_info is: build_info("5.27.0","2012-05-08 11:27:57","i686-pc-mingw32","GNU Common Lisp (GCL)","GCL 2.6.8").

Best regards,

Jean-Yves

----------------------------------------------------------------------

>Comment By: Jean-Yves (jyoberle)
Date: 2012-09-01 15:24

Message:
  Aleksasd,

  I agree with you regarding your calculations. However, when I apply the
function "trigsolve" to the equation, I get:

    trigsolve(cos(x)-sin(3*x),-%pi,%pi);
    {-(7*%pi)/8,-(3*%pi)/8,%pi/8,(5*%pi)/8}

  Several solutions are missing (e.g. %pi/4) because of the bug I
highlighted.

  Best regards,

  Jean-Yves


----------------------------------------------------------------------

Comment By: Aleksas (aleksasd)
Date: 2012-08-31 08:03

Message:
 To finding all solutions of trigonometric equation eq
from interval [a, b] we define function "trigsolve":

(%i1) trigsolve(eq,a,b):=block([s,i,ats,algebraic],
algebraic:true,
to_poly_solve([eq], [x],'simpfuncs =
['rootscontract,'expand,'radcan,'nicedummies]),
s:makelist(rhs(part(%%,k)[1]),k,1,length(%%)),
ats:[], 
for i:1 thru length(s) do 
(makelist(ev(s[i],%z0=k),k,-10,10),
ats:append(ats,%%)),
sublist(ats,lambda([e],e>=a and e<=b and 
float(ev(abs(lhs(eq)-rhs(eq)),x=e))<ratepsilon)),
sort(%%), setify(%%)
)$

 Example: solve cos(x)-sin(3*x)=0

(%i2) eq:cos(x)-sin(3*x)=0$
(%i3) cos(x)-cos(y)=-2*sin(1/2*x+1/2*y)*sin(1/2*x-1/2*y)$
(%i4) subst(y=3*x-%pi/2,%),expand;
(%o4) cos(x)-sin(3*x)=2*sin(x-%pi/4)*sin(2*x-%pi/4)
(%i5) eq1:sin(x-%pi/4)=0$
(%i6) eq2:sin(2*x-%pi/4)=0$
(%i7) S1:trigsolve(eq1,-%pi,%pi);
to_poly_solve: to_poly_solver.mac is obsolete; I'm loading
to_poly_solve.mac instead.
Loading maxima-grobner $Revision: 1.6 $ $Date: 2009-06-02 07:49:49 $
(%o7) {-(3*%pi)/4,%pi/4}
(%i8) S2:trigsolve(eq2,-%pi,%pi);
(%o8) {-(7*%pi)/8,-(3*%pi)/8,%pi/8,(5*%pi)/8}
(%i9) S:union(S1,S2);
(%o9) {-(7*%pi)/8,-(3*%pi)/4,-(3*%pi)/8,%pi/8,%pi/4,(5*%pi)/8}
(%i10) float(%), numer;
(%o10)
{-2.748893571891069,-2.356194490192345,-1.178097245096172,0.39269908169872,0.78539816339745,1.963495408493621}

 Answer: x=a+2*%pi*k, where a - any from S, k - any integer

(%i11) plot2d([cos(x)-sin(3*x)], [x,-%pi,%pi])$

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=104933&aid=3558096&group_id=4933