Re: Solve bug in Mathematica 5
- To: mathgroup at smc.vnet.net
- Subject: [mg44021] Re: [mg44010] Solve bug in Mathematica 5
- From: Andrzej Kozlowski <akoz at mimuw.edu.pl>
- Date: Sat, 18 Oct 2003 03:12:14 -0400 (EDT)
- Sender: owner-wri-mathgroup at wolfram.com
On Friday, October 17, 2003, at 06:14 PM, Artūras Acus wrote:
> Hi,
>
> The following example demonstrates Mathematica 5.0 "improvements"
> in Solve.
>
> lygtys2={E^(a*k*Cot[a*k])*Subscript[A, 1] - Cos[a*k]*Subscript[A, 2] +
> Sin[a*k]*Subscript[B, 2] == 0,
> -(E^(a*k*Cot[a*k])*k*Cot[a*k]*Subscript[A, 1]) -
> k*Sin[a*k]*Subscript[A, 2] -
> k*Cos[a*k]*Subscript[B, 2] == 0, Cos[a*k]*Subscript[A, 2] +
> Sin[a*k]*Subscript[B, 2] -
> E^(a*k*Cot[a*k])*Subscript[B, 3] == 0,
> -(k*Sin[a*k]*Subscript[A, 2]) + k*Cos[a*k]*Subscript[B, 2] -
> E^(a*k*Cot[a*k])*k*Cot[a*k]*Subscript[B, 3] == 0}
>
> Solve[lygtys2, {Subscript[A, 2], Subscript[B, 2], Subscript[B, 3],
> Subscript[A, 1]}]
>
> gives {0,0,0,0}, thought equations are lineary dependent. 4.1 returns
> correct rezult.
>
>
>
If you want to see the real improvement, look at the answer you get
with Reduce in v. 5 and compare it with the one version 4 returns.
I agree that Solve in v.5 should do a little better with this example
but actually it will give you the solution you want if you use the
Rational mode instead of the default Generic:
Solve[lygtys2, {Subscript[A, 2], Subscript[B, 2],
Subscript[B, 3], Subscript[A, 1]}, Mode -> Rational]
(long output suppressed).
Andrzej Kozlowski
Yokohama, Japan
http://www.mimuw.edu.pl/~akoz/
http://platon.c.u-tokyo.ac.jp/andrzej/