- #1
jimmy7430
- 2
- 0
Let [itex]Pr(X_{i} = +1) =\frac{2}{3} = 1 - Pr(X_{i} = -1) [/itex], and [itex] S_{n} = \sum{X_{i}}[/itex], For each k≥1, define [itex]T_{k}\ =\ min \left\{n≥1: S_{n} = k \right\}[/itex]. Calculate [itex]E[T_k][/itex], and [itex]Var[T_k][/itex].
2. Homework Equations
[itex]E[T_k] = \sum_{n=1}^{n=∞}{n Pr(T_k = n)} [/itex], and [itex]E[T_k]^2 = \sum_{n=1}^{n=∞}{n^2 Pr(T_k = n)} [/itex] will be used.
k is given as an known constant. A theorem says [itex] Pr(T_k = n ) = Pr(S_{1} S_{2} ... S_{n} ≠ 0, S_{n} = k) = \frac{|b|}{n}Pr(S_n = k)[/itex], and
[itex] Pr(S_n = k) = \left( \stackrel{n}{\frac{1}{2}(n+k)} \right) p^{\frac{1}{2}(n+k)}q^{\frac{1}{2}(n-k)} [/itex]
IMHO, problem should be asolved. However, It doesn't seem to have a nice expression. My guess will be use Sterling approximation, or Mathematica. Any comments are welcome.
2. Homework Equations
[itex]E[T_k] = \sum_{n=1}^{n=∞}{n Pr(T_k = n)} [/itex], and [itex]E[T_k]^2 = \sum_{n=1}^{n=∞}{n^2 Pr(T_k = n)} [/itex] will be used.
The Attempt at a Solution
k is given as an known constant. A theorem says [itex] Pr(T_k = n ) = Pr(S_{1} S_{2} ... S_{n} ≠ 0, S_{n} = k) = \frac{|b|}{n}Pr(S_n = k)[/itex], and
[itex] Pr(S_n = k) = \left( \stackrel{n}{\frac{1}{2}(n+k)} \right) p^{\frac{1}{2}(n+k)}q^{\frac{1}{2}(n-k)} [/itex]
IMHO, problem should be asolved. However, It doesn't seem to have a nice expression. My guess will be use Sterling approximation, or Mathematica. Any comments are welcome.