The Student Room Group

Calculating Life expectancy

Stuck with the following problem:

The Probability that a newborn will die at or before time t is given by the expression:

F0(t)=1(1t/100)1/4(t)=1-(1-{t/100})^{1/4}

0\le{t}\le{100}


Calculate the life expectancy of a person aged 40. How I've started:

Let Tx be the time of death of a person aged x.

Pr(TPr(Txt)=F\le{t}) = Fx(t)(t)

Let Pr(TPr(Tx>t)=S>{t}) = Sx(t)=1F(t)=1-Fx(t)(t)

Now Life expectancy (Expected number of years that x will live)
E(TE(Tx)=(tf) = \int(tfx(t))dt(t)) dt where ffx(t)(t) is the density function of F.

Integration by parts gives:

E(TE(Tx)=(S) = \int(Sx(t))dt(t)) dt

To find SSx(t)(t):

Pr(TPr(Tx>(t))=Pr(T>(t)) = Pr(To>x+tT>x+t\Bigm|To>x)>x)

Using the law of conditional Probability:

Pr(TPr(Tx>(t))=(Pr(T>(t)) = (Pr(To>x+t))/(Pr(T>x+t))/(Pr(To>x))>x))

Pr(TPr(Tx>(t))=S>(t)) = So(x+t)/S(x+t)/So(x)(x)

SSo(x)=(1x/100)1/4(x) = (1-{x/100})^{1/4}

So now to calculate Life expectancy,

E(TE(Tx)=(1+(t/(x100)))1/4dt) = \int (1+({t/({x-100)}}))^{1/4}dt

In our case, x = 40 and the integral is defined from 0 to 100, so

E(TE(T40)=0100(1+(t/60))1/4dt) = \int_0^{100} (1+({t/-60}))^{1/4}dt

I've integrated this to get:

4/5(x100)((t+x100)/(x100))5/4{4/5}(x-100)((t+x-100)/(x-100))^{5/4}

However when subbing in 100 to find the definite integral, you end up having to calculate (-2/3)^({5/4}). What do I do?
(edited 10 years ago)
Reply 1
bump

Quick Reply

Latest