On Mon, Feb 29, 2016 at 2:05 AM, AMiT Kumar <[email protected]> wrote:
> Hi Kshitij,
>
> Good to hear that you want to work on solvers. Here are my comments on your
> queries.
>
> 1. Search Based Solvers:
>
> This idea is currently vague, we need to figure out how exactly
> we can develop this. It would be ambitious to develop this idea and It
> should be
> accompanied with lot of research. Currently an abstract is written in the
> docs:
> https://github.com/sympy/sympy/blob/master/doc/src/modules/solvers/solveset.rst
>
> 2. Simplifying solutions returned from equations involving trigonometric
> expressions:
>
> With the Introduction of solveset an important problem of representing
> infinite solution
> has been solved to some extent, but still there are some issues with
> ImageSet Union,
> due to which a lot of simpler results are not displayed properly. This needs
> to be figured
> out to get better ImageSet Union.
>
> In [10]: solveset(sin(x), x)
> Out[10]: {2⋅n⋅π | n ∊ ℤ} ∪ {2⋅n⋅π + π | n ∊ ℤ}
>
> In [11]: solveset(sin(3*x), x)
> Out[11]:
>                                         ⎧        2⋅π        ⎫   ⎧        2⋅π
> {2⋅n⋅π | n ∊ ℤ} ∪ {2⋅n⋅π + π | n ∊ ℤ} ∪ ⎨2⋅n⋅π - ─── | n ∊ ℤ⎬ ∪ ⎨2⋅n⋅π + ───
> |
>                                         ⎩         3         ⎭   ⎩         3
>
>       ⎫   ⎧        π        ⎫   ⎧        π        ⎫
>  n ∊ ℤ⎬ ∪ ⎨2⋅n⋅π - ─ | n ∊ ℤ⎬ ∪ ⎨2⋅n⋅π + ─ | n ∊ ℤ⎬
>       ⎭   ⎩        3        ⎭   ⎩        3        ⎭
>
>
> Here is a previous approach on fixing this:
> https://github.com/sympy/sympy/pull/7673

Also https://github.com/sympy/sympy/pull/9500/files#r39220151.

Aaron Meurer

>
> 3. Implementing more Equation solvers
>
> A lot of thing which needs to be done is already done in old solve, like the
> solving of multivariate equation solver, you need to figure out how that
> works
> and how you can port those in solveset following the principles of solveset.
> https://github.com/sympy/sympy/issues/10008
>
>> I found this paper [5] which talks about implementing a parallel Gauss
>> method for solving this issue. Is it relevant ?
>
>
> This paper seems like implementing linear system solver, which is already
> implemented as linsolve.
>
> 4.  Solving f(x + a) - f(x) = 0 equations: [7]
>
> This needs to figured out.
>
>
>>
>>    5. Building the set infrastructure:
>> Implementing functions to handle multidimensional ImageSet
>>       Can we be more elaborate on what other features are we expecting ?
>
>
>
>  See this for motivation: https://github.com/sympy/sympy/issues/10008
>
>
> Best Regards,
> Amit Kumar
>
> On Tuesday, February 23, 2016 at 7:56:31 PM UTC+5:30, Kshitij Saraogi wrote:
>>
>> Hello,
>>
>>
>> I am Kshitij Saraogi and I will be a GSoC applicant this year under SymPy.
>>
>>
>> I wanted to discuss about the Solvers project.[1]
>>
>> After going through the discussions mentioned on the Ideas page, I would
>> like to get inputs on a few ideas for this project:
>>
>>
>>   1. Search based Solvers [2]:
>>
>>       I find this idea quite intriguing. I understand only an abstract
>> view of the idea was presented there.
>>
>>       I think we should try to implement this as it would make the API
>> cleaner and robust.
>>
>>       Since, not much has been written about this, I would like to know
>> more about it.
>>
>>    - What is the methodology we are thinking to use for ranking solutions
>> (if any)?
>>    - What would be the parameters on which the cost function of different
>> sets depend ?
>>       So, I would appreciate guidance in this direction.
>>
>>
>>    2.  Simplifying solutions returned from equations involving
>> trigonometric expressions:
>>
>>         The solveset module needs improvement with regards to the
>> trignometric equation solver.
>>
>>         An equation,as simple as, sin(x)=0 gives an output which should be
>> simplified.
>>
>>         This should be a big concern.
>>
>>         What would be some good starting points to get an overview of the
>> issue and possibly a few ideas to resolve this ?
>>
>>
>>    3. Implementing more equation solvers: [3]
>>
>> System of  multivariate linear equations.
>>
>> Nonlinear multivariate equation solver.
>>
>> Equations solvable by LamberW function (Transcendental equation solver)
>>
>> Nested trignometric expressions.
>>
>>  As Amit pointed out here[4], that we need (ii) and (iii) to make solveset
>> at par with solve.
>>
>>          I found this paper [5] which talks about implementing a parallel
>> Gauss method for solving this issue. Is it relevant ?
>>
>>          While fixing an issue [6], I came to know that we need a more
>> reliable multivariate nonlinear equation solver.
>>
>>          I would like to know more about these solvers with respect to
>> their immediate need and the possible methods of their implementation.
>>
>>
>>
>>    4.  Solving f(x + a) - f(x) = 0 equations: [7]
>>
>>          While going through some issues, I found that the current solvers
>> can’t handle these type of equations.
>>
>>
>>
>>    5. Building the set infrastructure:
>>
>> Implementing functions to handle multidimensional ImageSet
>>
>>       Can we be more elaborate on what other features are we expecting ?
>>
>> I would really appreciate if someone can point out the issues I may have
>> missed.
>>
>> Also, any relevant resources or links for further readings would help too.
>>
>>
>> Thanks,
>>
>> Kshitij Saraogi
>
> --
> You received this message because you are subscribed to the Google Groups
> "sympy" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to [email protected].
> To post to this group, send email to [email protected].
> Visit this group at https://groups.google.com/group/sympy.
> To view this discussion on the web visit
> https://groups.google.com/d/msgid/sympy/98d1cacd-0ff4-45ef-a0b7-d4b1b2a47c28%40googlegroups.com.
>
> For more options, visit https://groups.google.com/d/optout.

-- 
You received this message because you are subscribed to the Google Groups 
"sympy" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
To post to this group, send email to [email protected].
Visit this group at https://groups.google.com/group/sympy.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/sympy/CAKgW%3D6K7Na9DLYFOfyibELiYGGcNdqAEeEY5fhE0847UV7yayA%40mail.gmail.com.
For more options, visit https://groups.google.com/d/optout.

Reply via email to