Thread Links Date Links
Thread Prev Thread Next Thread Index Date Prev Date Next Date Index

Re: floats and intervals, constants, 0.1, 1e400 and Inf



On 12 Nov 2008 at 8:42, Arnold Neumaier wrote:


> Chenyi Hu schrieb:
>  > We have applied midpoint-radius representations in recent work.
>  > It certainly has its advantage in handling uncertainty.
> 
> Could you please state the computational advantages compared
> to the inf-sup representation rather than only asserting
> their existence?
> 

An important computational advantage could be the usage
of 1-2 digit precision of the radius. For a motivation see:

Van Emden, M. H., On the Significance of Digits in
Interval Notation, Reliable Computing 10, 1, 45–58.
 

> Clearly the representation is useful for interpreting data and results 
> of an application. But this only affects conversion from and to text,
> not the arithmetic with mid-rad intervals.

On the contrary, I argue that the MR-arithmetic is simpler that the
inf-sup one. Moreover, it is valid for Kaucher/modal intervals without
any additional cost.

Svetoslav

> 
> Which package do you use for mid-rad interval arithmetic?
> 
> 
> Arnold Neumaier
> 
> 
> 
>  >
>  >>>> "Svetoslav Markov" <smarkov@xxxxxxxxxx> 11/11/2008 2:04 AM >>>
>  > On 10 Nov 2008 at 13:02, Arnold Neumaier wrote:
>  >
>  >> Though briefly discussed, I think nobody has shown strong emotions for
>  >> satandardizing mdpoint-radius interval representations.
>  >> I advocate against it.
>  >>
>  >
>  > I should like to declare myself as a strong supporter of
>  > midpoint-radius interval representations.
>  >
>  >  Svetoslav Markov
>  >
>  >