Opened 2 months ago

Last modified 4 weeks ago

#27314 new defect

Can't solve symbolic equations containing inexact (e.g. RIF, RBF) numbers

Reported by: rburing Owned by:
Priority: major Milestone: sage-8.8
Component: symbolics Keywords: RIF, RBF, solve
Cc: Merged in:
Authors: Reviewers:
Report Upstream: N/A Work issues:
Branch: Commit:
Dependencies: Stopgaps:


There are (undocumented?) symbolic wrappers around RIF, RBF objects, but solving equations with them

sage: var('x')
sage: solve(x + RIF(0.999,1.001) == 0, x)

fails because the inexact objects cannot be converted to Maxima objects.

This is confusing to new users; see Ask SageMath: long traceback when calling solve().

Specifying algorithm='sympy' gives a solution but loses information about precision.

The least thing that could be done is for solve(eqn, x) to check eqn.is_exact() and give a warning.

Change History (2)

comment:1 Changed 2 months ago by charpent

One notes that numerical root seeking via find_root also fails for both RIF and RBF :

sage: u=RIF(0.99,1.01);u
sage: s=find_root(x+u,-3,3);s
sage: type(s)
<type 'float'>
sage: v=RBF(v);v
sage: v=v.add_error(0.01);v
[1.0 +/- 0.0101]
sage: s2=find_root(x+v,-3,3);s2
sage: type(s2)
<type 'float'>

Extending find_root to handle RIF and RBF expressions would allow to handle the cases where no explicit form of a root solution is known (e. g. implicit expressions resulting of solve), where find_root allows to compute a numerical approxomation.

This would certainly be handy (but is probably a significant project in itself !).

It might also be useful to check which of the optimization functions in Sage support such uncertainty handling...

Last edited 2 months ago by charpent (previous) (diff)

comment:2 Changed 4 weeks ago by embray

  • Milestone changed from sage-8.7 to sage-8.8

Ticket retargeted after milestone closed (if you don't believe this ticket is appropriate for the Sage 8.8 release please retarget manually)

Note: See TracTickets for help on using tickets.