If you mean "problems that can be transformed
to polynomials in one variable that can be solved
exactly" then I think a half page of code could do
that. Find zeros. for each interval between zeros
determine if polynomial is positive or negative.
This might be a useful start, but how far do you
want to push this?
RJF
Gosse Michel wrote:
> Hello
>
> I think something like :
>
> solve(2*x+5>0);
> must return x>-5/2 or realrange(-5/2,inf)
>
> solve(x^2+3x*-4>0)
> must return x<-4 or x>1
>
> solve((x+1)/(x+3)>5)
>
> I think maple can do that.
>
> Best regards
>
>
>>Can you define what you mean by solving inequalities and
>>what algorithms you would expect?
>>
>>For example, given the range of possible expressions
>>in maxima, the predicate "f(x)>0" is undecidable.
>>On the other hand, is(3<4) is probably already solved.
>>
>>Real inequalities? integer inequalities?
>>
>>I don't know if anyone has plans to do better, but
>> it is certainly possible to add such features to
>>maxima. If you volunteer to do so, I think people
>>would be glad to give you advice.
>>
>>is (abs(sin(x))<=1) for example? The commercial macsyma says true.
>>
>>RJF
>>
>>Gosse Michel wrote:
>>
>>>Hello
>>>
>>>Is it planned to add solving inequalities for the future release of
>>>maxima ?
>>>
>>_______________________________________________
>>Maxima mailing list
>>Maxima@www.math.utexas.edu
>>http://www.math.utexas.edu/mailman/listinfo/maxima
>>
>