That's odd

Suppose two real numbers x x and y y are chosen, uniformly and at random, from the open interval ( 0 , 1 ) (0,1) . Let P P be the probability that the integer closest to x y \dfrac{x}{y} is odd.

Find 10000 P \lceil 10000P \rceil .


The answer is 5354.

This section requires Javascript.
You are seeing this because something didn't load right. We suggest you, (a) try refreshing the page, (b) enabling javascript if it is disabled on your browser and, finally, (c) loading the non-javascript version of this page . We're sorry about the hassle.

2 solutions

Let x x be chosen, uniformly and at random, from the interval ( 0 , 1 ) (0,1) on the x x -axis, and y y from the interval ( 0 , 1 ) (0,1) along the y y -axis. So the sample space is a unit square with lower-left corner ( 0 , 0 ) (0,0) and upper-right corner ( 1 , 1 ) (1,1) , (the sides of which will be represented by dotted lines, so as to avoid division by zero).

Then P P will be the combined areas of the regions within this square where the closest integer to x y \dfrac{x}{y} is odd. This condition can be quantified as saying that

x y ( 2 n + 1 ) < 1 2 ( 2 n + 1 2 ) < x y < ( 2 n + 3 2 ) | \dfrac{x}{y} - (2n + 1)| \lt \dfrac{1}{2} \Longrightarrow (2n + \dfrac{1}{2}) \lt \dfrac{x}{y} \lt (2n + \dfrac{3}{2})

for some non-negative integer n n . For a given value n n the corresponding region within the unit square will be that lying between the lines

y = 2 4 n + 1 x y = \dfrac{2}{4n + 1}*x and y = 2 4 n + 3 x y = \dfrac{2}{4n + 3}*x .

For n = 0 n = 0 the area of the corresponding region is 1 4 + 1 6 \dfrac{1}{4} + \dfrac{1}{6} , and for n 1 n \ge 1 the area is 1 4 n + 1 1 4 n + 3 \dfrac{1}{4n + 1} - \dfrac{1}{4n + 3} . we thus have that

P = 1 4 + 1 6 + 1 5 1 7 + 1 9 1 11 + . . . . . = P = \dfrac{1}{4} + \dfrac{1}{6} + \dfrac{1}{5} - \dfrac{1}{7} + \dfrac{1}{9} - \dfrac{1}{11} + ..... =

5 12 1 + 1 3 + n = 0 ( 1 ) n 2 n + 1 = 1 4 + π 4 = 0.53539816..... \dfrac{5}{12} - 1 + \dfrac{1}{3} + \displaystyle\sum_{n=0}^{\infty}\dfrac{(-1)^{n}}{2n + 1} = -\dfrac{1}{4} + \dfrac{\pi}{4} = 0.53539816..... .

Thus 10000 P = 5354 \lceil 10000*P \rceil = \boxed{5354} .

Alternatively, extend the quadrilateral that you find the area of, until it is a triangle (with the same height as all the other triangles, and the base is on the same line as all the other triangles).

The area of all the triangles is 1 1 3 + 1 5 = π 4 1-\dfrac{1}{3}+\dfrac{1}{5}-\cdots = \dfrac{\pi}{4}

Now we just subtract the area of the extended part, which is 1 4 \dfrac{1}{4} so our total area is π 1 4 \dfrac{\pi-1}{4} as you found.

Daniel Liu - 6 years, 4 months ago

sir Sorry , But I didn't understand why m o d ( x / y ( 2 n + 1 ) ) < 1 / 2 mod(x/y-(2n+1))<1/2 ?

Karan Shekhawat - 6 years, 4 months ago

Log in to reply

Those are absolute value signs around x / y ( 2 n + 1 ) x/y - (2n + 1) . So this just means that x / y x/y is within a "radius" of 1 / 2 1/2 of an odd integer 2 n + 1 2n + 1 .

Brian Charlesworth - 6 years, 4 months ago

Why is this bounded by a unit square when the values of x/y can approach infinity?

Greg Grapsas - 4 years, 9 months ago

Log in to reply

The unit square in question doesn't include the boundary, i.e., the two axes, so as to avoid the division by zero issue. For coordinates ( x , y ) (x,y) within this grid, though, we are evaluating the nearest integer to f ( x , y ) = x y f(x,y) = \dfrac{x}{y} , which is well-defined since y = 0 y = 0 is not included in the domain.

Brian Charlesworth - 4 years, 9 months ago
Soumava Pal
Mar 4, 2016

This problem came in Putnam 1993 . And can be done once we know that p i / 4 pi/4 =1-1/3+1/5-1/7+1/9-....

Can identity π 4 = 1 1 / 3 + 1 / 5 1 / 7 + . . . \frac{\pi}{4}=1-1/3+1/5-1/7+... proof by using taylor series? I use l n ( x + 1 ) = x x 2 2 + x 3 3 x 4 4 + x 5 5 . . . ln(x+1) = x-\frac{x^2}{2}+\frac{x^3}{3}-\frac{x^4}{4}+\frac{x^5}{5}-... and compute x = i x=i , but isn't the series can only used when x x is near 0? How can I say i i is near to 0? i i isn't a real number! Although I used it and got the right answer, but I can't understand it. Or this identity can be proof using another method?

Kelvin Hong - 3 years, 9 months ago

0 pending reports

×

Problem Loading...

Note Loading...

Set Loading...