B2.12

Probability and Measure
Part II, 2001

(a) Let Ω=(0,1),F=B((0,1))\Omega=(0,1), \mathcal{F}=\mathcal{B}((0,1)) be the Borel σ\sigma-field and let P\mathbb{P} be Lebesgue measure on (Ω,F)(\Omega, \mathcal{F}). What is the distribution of the random variable ZZ, where Z(ω)=2ω1Z(\omega)=2 \omega-1 ?

Let ω=n=12nRn(ω)\omega=\sum_{n=1}^{\infty} 2^{-n} R_{n}(\omega) be the binary expansion of the point ωΩ\omega \in \Omega and set U(ω)=n odd 2nQn(ω)U(\omega)=\sum_{n \text { odd }} 2^{-n} Q_{n}(\omega), where Qn(ω)=2Rn(ω)1Q_{n}(\omega)=2 R_{n}(\omega)-1. Find a random variable VV independent of UU such that UU and VV are identically distributed and U+12VU+\frac{1}{2} V is uniformly distributed on (1,1)(-1,1).

(b) Now suppose that on some probability triple XX and YY are independent, identicallydistributed random variables such that X+12YX+\frac{1}{2} Y is uniformly distributed on (1,1)(-1,1).

Let ϕ\phi be the characteristic function of XX. Calculate ϕ(t)/ϕ(t/4)\phi(t) / \phi(t / 4). Show that the distribution of XX must be the same as the distribution of the random variable UU in (a).