achille
Posts:
575
Registered:
2/10/09


Re: Probability Pill
Posted:
Dec 31, 2012 7:16 AM


On Thursday, December 27, 2012 5:39:14 PM UTC+8, William Elliot wrote: > Each day I take 1/2 an aspirin tablet. I bought a bottle of 100 tablets; > > each day I take out one, if it's whole I break it half and eat a half and > > put the other half back: if I pull out a half tablet I eat it. I was > > wondering after I break the last whole one what the expected number of > > halves are in the bottle? I assume that any piece I pull out has uniform > > probability.
Let f_{m,n} be the expected value of final number of half pills if one start with m full and n half pills. f_{m,n} satisfies:
[1a] (m+n) f_{m,n} = m f_{m1,n+1} + n f_{m,n1} for m > 0 [1b] f_{0,n} = n
Let f_m(t) be \sum_{n=0..oo} f_{m,n} t^n, [1*] is equivalent to:
[2a] ( m  t + t*(1t)*d/dt ) f_m(t) = (m/t)*(f_{m1}(t)f_{m1}(0)) for m > 0. [2b] f_0(t) = t/(1t)^2
Let u = t/(1t) and g_m(t) = (1t) f_m(t), one can simplify [2*] to
[3a] ( m + u*d/du ) g_m(u) = (m/u)*((1+u)*g_{m1}(u)  g_{m1}(0)) for m > 0. [3b] g_0(u) = u
For m > 0, on the L.H.S of [3a],
( m + u*d/du) is an invertible operator over the spaces of polynomials. furthermore, this operator preserve the degree of a polynomial.
on the R.H.S of [3a],
if g(u) is a polynomial of degree k, so does (m/u) * ((1+u)g(u)  g(0)). Since [3b] says g_0(u) is a linear polynomial, these observations implies all g_m(t) are linear polynomials.
Let g_m(u) = a_m + b_m*u and substitute these in [3*], we get:
[4a] a_m = a_{m1} + b_{m1} (m+1) b_m = m b_{m1} [4b] a_0 = 0 b_0 = 1 From these, we get
b_m = 1/(m+1) and a_m = a_0 + \sum_{k=0..m1} b_{k} = \sum_{k=1..m} 1/m. So the number we want f_m,0 = f_m(0) = g_m(0) = a_m = 1 + 1/2 + .. + 1/m as quasi has pointed out in a previous post.
the ( m + u*d/du ) in the L.H.S of [3a] is a invertible operator over the spaces of polynomials which also preserve the degree of polynomial. Furthermore, if g(u) is a polynomial of degree k, so does the sort of expression (m/u) * (( 1 + u)g(u)  g(0) appears in the R.H.S of [3b]. These observations together with g_0(u) is a polynomial

