跳到主要内容

東京大学 情報理工学系研究科 コンピュータ科学専攻 2017年8月実施 専門科目II 問題6

Author

祭音Myyura (co-authored with GPT 5.6 SOL)

Description

Consider the decomposition time of an RNA molecule. Assume that the probability density function of the decomposition time TT is

fT(t)=λeλt,t0,f_T(t)=\lambda e^{-\lambda t},\qquad t\ge0,

where λ\lambda is a positive real constant.

Answer the following questions.

(1) Calculate the cumulative distribution function

FT(t)=0tfT(x)dx.F_T(t)=\int_0^t f_T(x)\,dx.

Also compute the median of TT.

(2) We measured the decomposition times TiT_i (i=1,,ni=1,\ldots,n) of nn RNA molecules. Assume that the decomposition time of each RNA molecule follows the probability density function fT(t)f_T(t) independently and identically. Calculate the expected value and the variance of

μT=i=1nTin.\mu_T=\frac{\sum_{i=1}^nT_i}{n}.

(3) Let Tmax=max{T1,,Tn}T_{\max}=\max\{T_1,\ldots,T_n\}, which is the maximum of the measured times TiT_i in question (2). Let Prob(Tmax>t)\operatorname{Prob}(T_{\max}>t) denote the probability that Tmax>tT_{\max}>t. Give an expression for Prob(Tmax>t)\operatorname{Prob}(T_{\max}>t) in terms of FT(t)F_T(t).

(4) Calculate the probability density function fTmax(t)f_{T_{\max}}(t) of TmaxT_{\max}, and the expected value of TmaxT_{\max}.

题目描述

RNA 分子的分解时间 TT 的概率密度为

fT(t)=λeλt,t0,f_T(t)=\lambda e^{-\lambda t},\qquad t\ge0,

其中 λ>0\lambda>0

(1)求累积分布函数 FT(t)F_T(t)TT 的中位数。

(2)独立同分布地测量 nn 个 RNA 分子的分解时间 T1,,TnT_1,\ldots,T_n。求样本均值 μT=1niTi\mu_T=\frac1n\sum_iT_i 的期望和方差。

(3)令 Tmax=max(T1,,Tn)T_{\max}=\max(T_1,\ldots,T_n)。用 FT(t)F_T(t) 表示 Pr(Tmax>t)\Pr(T_{\max}>t)

(4)求 TmaxT_{\max} 的概率密度和期望。

Kai

(1)

t0t\ge0

FT(t)=0tλeλxdx=1eλt.F_T(t)=\int_0^t\lambda e^{-\lambda x}\,dx=1-e^{-\lambda t}.

t<0t<0 时,FT(t)=0F_T(t)=0

FT(m)=1/2F_T(m)=1/2,得中位数

m=ln2λ.m=\frac{\ln 2}{\lambda}.

(2)

指数分布满足 E[Ti]=1/λ\mathbb E[T_i]=1/\lambdaVar(Ti)=1/λ2\operatorname{Var}(T_i)=1/\lambda^2。由独立性,

E[μT]=1λ,Var(μT)=1nλ2.\mathbb E[\mu_T]=\frac1\lambda, \qquad \operatorname{Var}(\mu_T)=\frac1{n\lambda^2}.

(3)

t0t\ge0

Pr(Tmax>t)=1Pr(T1t,,Tnt)=1[FT(t)]n.\Pr(T_{\max}>t) =1-\Pr(T_1\le t,\ldots,T_n\le t) =1-[F_T(t)]^n.

(4)

FTmax(t)=[FT(t)]nF_{T_{\max}}(t)=[F_T(t)]^n

fTmax(t)=n[FT(t)]n1fT(t)=nλeλt(1eλt)n1,t0.f_{T_{\max}}(t) =n[F_T(t)]^{n-1}f_T(t) =n\lambda e^{-\lambda t}(1-e^{-\lambda t})^{n-1}, \qquad t\ge0.

再用尾积分公式,

E[Tmax]=0{1(1eλt)n}dt=1λk=1n1k.\begin{aligned} \mathbb E[T_{\max}] &=\int_0^\infty\{1-(1-e^{-\lambda t})^n\}\,dt\\ &=\frac1\lambda\sum_{k=1}^n\frac1k. \end{aligned}