Group Abstract Group Abstract

Message Boards Message Boards

Another inequality to solve

Attachments:
POSTED BY: Xavier Koch
4 Replies
Posted 5 years ago
POSTED BY: Bill Nelson

Actually I don't get how but I attached the wrong file ... This one is the good one ;)

Attachments:
POSTED BY: Xavier Koch
Posted 5 years ago

This is very slow

Condition34=Reduce[{(-a+c β+t θ1)^2(6b^2+4b λ+λ^2)/(8b^3)>
(3a^2-6a c α+3c^2 α^2+2a(-3+2p)t θ2+2c(3-2p)t α θ2+(3-4p+2p^2)t^2 θ2^2)/(4b),
a>c>0,1>p>0,1>t>0,1>b>0,1>λ>0,β>α>1,a>c α,a>c β,0<θ1<θ2<θ3<1},β]

Verify we can divide both sides

Simplify[(6b^2+4b λ+λ^2)/(8b^3)>0,
{a>c>0,1>p>0,1>t>0,1>b>0,1>λ>0,β>α>1,a>c α,a>c β,0<θ1<θ2<θ3<1}]

returns

True

Try to make the problem easier, but this is still very slow

Condition34=Reduce[{(-a+c β+t θ1)^2>
(3a^2-6a c α+3c^2 α^2+2a(-3+2p)t θ2+2c(3-2p)t α θ2+(3-4p+2p^2)t^2 θ2^2)/(4b)/(6b^2+4b λ+λ^2)*(8b^3),
a>c>0,1>p>0,1>t>0,1>b>0,1>λ>0,β>α>1,a>c α,a>c β,0<θ1<θ2<θ3<1},β]

Try to make the problem easier

Condition34=Reduce[{(-a+c β+t θ1)^2>q,
a>c>0,1>p>0,1>t>0,1>b>0,1>λ>0,β>α>1,a>c α,a>c β,0<θ1<θ2<θ3<1},β]/.
q->(3a^2-6a c α+3c^2 α^2+2a(-3+2p)t θ2+2c(3-2p)t α θ2+(3-4p+2p^2)t^2 θ2^2)/(4b)/(6b^2+4b λ+λ^2)*(8b^3)

The output is very large because we have not applied any domain information

This is still slow, but it may provide a more compact result for you

Simplify[Condition34,{a>c>0,1>p>0,1>t>0,1>b>0,1>λ>0,β>α>1,a>c α,a>c β,0<θ1<θ2<θ3<1}]

This

(Reduce[{(-a+c β+t θ1)^2>q,
a>c>0,1>p>0,1>t>0,1>b>0,1>λ>0,β>α>1,a>c α,a>c β,0<θ1<θ2<θ3<1},β]/.
q->(3a^2-6a c α+3c^2 α^2+2a(-3+2p)t θ2+2c(3-2p)t α θ2+(3-4p+2p^2)t^2 θ2^2)/(4b)/(6b^2+4b λ+λ^2)*(8b^3))/.
{Inequality[0, Less, c, Less, a]->True,Inequality[0, Less, p, Less, 1]->True,
Inequality[0, Less, t, Less, 1]->True,Inequality[0, Less, b, Less, 1]->True,
Inequality[0, Less, λ, Less, 1]->True,Inequality[0, Less, a, LessEqual, 1]->True,
Inequality[θ1, Less, θ2, Less, 1]->True,Inequality[θ2, Less, θ3, Less, 1]->True,
Inequality[α, Less, β, Less, a/c]->True,Inequality[0, Less, θ1, Less, 1]->True,
Inequality[1, Less, α, Less, a/c]->True
}

will be much faster than Simplify, but it will not make the result as simple and it is necessary to be very careful with the pattern matching to make certain the replacements are exactly correct. If you look at the result of this there will still be many inequalities that might be replaced with True and thus eliminated if it is possible to justify the replacements. The result is still very large, but it might be possible to use this method to get it to be small enough that Simplify could then be done in acceptable time.

Please check all this very carefully.

POSTED BY: Bill Nelson
POSTED BY: EDITORIAL BOARD
Reply to this discussion
Community posts can be styled and formatted using the Markdown syntax.
Reply Preview
Attachments
Remove
or Discard