Re: Simplification: science or heuristics?

carlos_at_colorado.edu
Date: 01/20/05 

Date: 20 Jan 2005 10:43:48 -0800

gmatthew@amgen.com wrote:
> Well the first so called simplification of
>
> 1-x^2
> ----- -> 1 + x
> 1-x
> isn't even correct, since the two expressions have different domains.

This is largely pointless and off topic. In a well designed ES system,
there is a symbol database that returns information on the domain and
range of each symbol, or uses default assumptions such as "symbols
without declared assumptions are assumed to be complex and
unrestricted" Thus x*y-y*x -> 0 is not correct if x and y are members
of a noncommutative ring.

Relevant Pages

  • Re: remove negative indices
    ... does multiplying the numerator and denominator by y^2 do it? ... spot checking could lead you astray. ... that the simplification is valid. ... for some types of expressions (for example ...
    (sci.math)
  • Re: Simplification: science or heuristics?
    ... since the two expressions have different domains. ... the simplification is also perfectly correct, ... very specific interpretation and the notion of being “incorrect” is also ... arithmetics or arithmetics over sets, and even a-a = 0 is no longer valid. ...
    (sci.math.symbolic)
  • Re: Improving the MATLAB function "simple" / genetic not a good idea
    ... trig identities. ... If you want to improve the simplifier for "all expressions" then there should be better methods, ... A genetic algorithm, it seems to me, has very little chance of finding a good simplification sequence for an expression ... Of course some computer algebra systems also include rules; used with restraint they can put the "finishing touches" on an expression. ...
    (sci.math.symbolic)
  • Re: Advanced expression simplification
    ... > and then applies various rules to the parsed tree. ... Much of it already works (reduction of constants, ... > If I do such things, I need to make sure that simplification does not ... Some approaches can initially lead to bigger expressions, ...
    (comp.compilers)
  • Re: Using exponents law indiscriminately in Calculus I
    ... Can someone with experience in Calc 1 and it's many manipulations of expressions using exponent laws tell me why we are safe in doing this? ... Sometimes the simplification will give you a function with a larger domain, ...
    (sci.math)