finding a number that satisfies given conditions

1 vue (au cours des 30 derniers jours)
Beaya
Beaya le 25 Oct 2013
I need to find a floating point number x from the interval 1 < x < 2 such that: x ∗ (1/x) isn't equal to 1. I'ma beginner at Matlab and don't know how this can be done. I was given a clue that using integrals could help in solving this. Could anyone help me?
  3 commentaires
Walter Roberson
Walter Roberson le 25 Oct 2013
Modifié(e) : Walter Roberson le 25 Oct 2013
Multiplication, I would say. Looks like an exercise in exploring the limits of binary floating point representation.
Beaya
Beaya le 1 Nov 2013
I wrote a program to do this in C, but I must have done something wrong because the result was: 4.94066e - 324, which is smaller than 1.

Connectez-vous pour commenter.

Réponses (2)

Mike Hosea
Mike Hosea le 4 Nov 2013
Integrals? Not sure about that. A random search will finish faster. There are LOTS. You can use LINSPACE to generate evenly spaced trial numbers, or you can us 1 + rand(1,n) to generate n random numbers between 1 and 2. Then you can check the condition using MATLAB's elementwise array operators: bool = x.*(1./x) ~= 1. Notice the .* instead of *, since * without the . is matrix multiplication. Same thing for division. Actually if you generate a boolean array that way from a set of trial numbers, you can also extract the results from your trial vector x using MATLAB's logical indexing, e.g. x(bool). Try it!

Walter Roberson
Walter Roberson le 4 Nov 2013
Do a binary search over the interval realmin() to 1, to find the boundary point.

Catégories

En savoir plus sur Logical dans Help Center et File Exchange

Tags

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by