Professional Documents
Culture Documents
7 MFQR II Chapter Five
7 MFQR II Chapter Five
MooBLS FoR
S unvrvar- C oNTTNGENT Rr srs
The second section of the text addresses traditional contingent payment models where pay-
ments are dependent on the continued survival (or failure to survive) of a defined entity. This
entire subject has traditionally been known as life contingencies, since the payment(s) were
contingent on the continued life (or death) of a designated person. In this textbook the sub-
ject is treated in a more general fashion, where "survival" can just as easily mean the contin-
uation of a labor strike, the functioning of a mechanical device (often called reliability
theory), or the regular payment of scheduled coupons under a corporate bond.
Discrete functions numerically evaluated from a discrete survival model (i.e., a life table) are
emphasized, since they represent most of the practical problems encountered by actuaries.
The more abstract continuous functions are also considered, frequently for their interesting
mathematical behavior. A random variable approach is employed along with a deterministic
approach when that view is useful to broaden the reader's understanding.
The survival model is presented in Chapters 5 and 6, in both its parametric and tabular con-
texts.
The standard set of single-life, single-decrement actuarial topics is covered in Chapters 7-11:
contingent payment models (with emphasis on their standard insurance applications), con-
tingent annuities (life annuities), annual funding schemes (annual premiums), including their
m'htv andcontinuous variations, and contingent contract reserves, now presented in two chap-
ters. Extensions to the multi-status cases of joint and last-survivor are presented in Chapter
12 andmultiple decrement models are covered in Chapters 13 and 14.
A discussion of using Microsoft Excel to calculate actuarial values from a life table is pre-
sented in Appendix A.
I
Altematively, the tabular model is also called
a mortality table.
' Another term for failure is- decremenl.lf an
entityl-ras a.particular stafus, such as survival,
that status is often described aiiig'i""ri."ri"i r;i^ then failure to retain
context of multiple decrements,^t *trriit we t{iffi;. ffiiliL#r:norogy is particurarry usetur in the
"n"o"r[i
i" brr"pi;;'il'";ji;:'"'
61
62 T CHAPTERFIVE
We assume that the entity exists at age 0, so the domain of the random variable fs is f0 > 0.
We refer to Zs as the age-at-failure randomvariable. We will consider the terms "failute" and
*death" to be synonymous, so we will also refer to Is as the age-at-deathrandom variable.3
It is easy to see that the numerical value of the age at failure is the same as the length of time
that sgrvival lasts until failure occurs, since the variable begins at age 0, so we can also refer
to 76 as the timeio-failure randomvariable. (If failure occurs at exact age t,then / is also the
time until failure occurs.)
Later (see Section 5.3) we will consider the case where the entity of interest is-known to
have survived to some age x>0. Then the time-to-failure random variable, to be denoted by
{,, will not be identical to the age-at-failure random variable 7e, although they will be re-
lated to each other by 4 = x+7". When dealing with this more general case we will do our
thinking in terms of the time-to-failure random variable.
for r > 0.4 we have already noted, however, that 'To = 0 is not possible' will always
so we
consider that F6(0) = 0. We observe that d(r) gives the probability that failure will occur
prior to (or at) precise age t for our entity known to exist at age 0. In actuarial notation, this
probability is denoted by ,qo, so we have
for I > 0. since we take f'o(0) = 0, it follows that wq will always take so(0) = 1' The SDF
gives the probability that the age at failure exceeds r, which is the same as the probability
ihat the ttror
to exist at age 0 will survive to age l. Since the notion of infinite sur-
"ntity
vival is unrealistic, we consider that
3
In practice, age-at-failure is often used for inanimate objects, such as light bulbs or labor strikes, and age-at-
death is used for animate entities, such as laboratory animals or human persons under an insufttnce arrangement.
a
In probability theory, it is customary to subscript the CDF symbol with the name of the random variable, which
suggests the notation F6(r) in this case. With the name of the random variable understood to be 7i, in this sec-
tion, we prefer the notation f0(/) to avoid the awkwardness of subscripting a subscript'
Sunvrv,c,L MoDELs | 63
lim,So(r) g
t-t6
= (s.aa)
and
limff(r) = 1. (5.4b)
t)6
In probability textbooks in general, the CDF is given greater emphasis than is the SDF. (Some
textbooks do not even define the SDF at all.) But when we are dealing with an age-at-failure
random variable, and its associated distribution, the SDF will receive greater attention.
EXAMPLE 5.I
Use both the CDF and the SDF to express the probability that an entity known to exist at age
0 will fail between the ages of 10 and 20.
Sor,rmoN
We seek the probability that Zo will take on a value between 10 and 20.Interms of the CDF
we have
Pr(t}<4 < 20) = FeQ0\-400).
fo(t) = 4ro@=
dt -!sogy,
dt
(s.6)
and
,so(r) = [,* fou)ar. (s.8)
li root a, - 1. (5.e)
It is important to recognize that /o (r) is the unconditional density of failure at age l. By this
we mean that it is the density of failure at age / given only that the entity existed at t=0.
The significance of this point will become clearer in the next subsection.
Recall that the PDF of foQ), is the unconditional density of failure at age /. We now de-
To,
frne a conditional density of failure at age l, with such density conditional on survival to age
r. This conditional instantaneous measure of failure at age /, given survival to age r, is called
the hazard rate at age t, or the hazard rate functio,n (HRF) when viewed as a function of r.
(In some textbooks the hazard rate is called thefailure rate.)ltwill be denoted bV h@.
hQ):# (s.1 1)
Equations (5.11) and (5.6) give formal defuritions of the HRF and the PDF, respectively, of the
age-at-failure random variable. Along with the definitions it is also important to have a clear
understanding of the conceptual meanings of )"(t) and fi(r). They are both instantaneous
measures of the density of failure at age t; they differ from each other in that 4(r) is condi-
tional on survival to age /, whereas /r(l) is unconditional (i.e., given only existence at age 0).
In the actuarial context of survival models for animate objects, including human persons, failure
means death, or mortality, and the hazardrate is normally called theforce of mortality. We will
discuss the actuarial context further in Section 5.1.6 and in Chapter 6.
Sur.irryarMoDEls r 65
Some important mathematical consequences follow directly from Equation (5.11). Since
fo Q) = - * t oe),it follows that
-frd soro d
hG) =
roft) -
= -
-alnJo(t)'
Ins^(r)- (5'12)
Integrating, we have
Exauplu 5.2
for 0<t< 100. Find (a) the PDF and (b) the HRF forthis random variable.
Sor,wroN
otlllotl,;'"' .50(100-r)-r.
v\ /
hQ) = [:\,! =
so0) .lo(1oo-r)ti2= ---\-- o
66 r CHAPTERFTvE
provided
l$t.So(r) = 0, a form which is frequently used to find the first moment of an age-
at-failure random variable.
Specific expressions can be developed for the moments of ?i for specific forms of S(r).
This will be pursued in the following section.
Another property of the age-at-failure random variable that is of interest is its median value.
We recall that the median of a continuous random variable is the value for which there is a
507o chance that the random variable will exceed (and thus also not exceed) that value.
Mathematically,y is the median of 7i, if
I
Pr(To>Y) = Pr(To3Y) (s.21)
2'
When the age-at-failure random variable is considered in an achrarial context, special sym-
bols are used for some of the concepts defined in this section. The hazard rate, now called
the force of mortality, is denoted by F,, rather than )oQ). Thus we have
SunvrvelMoosls . 67
,,=-4#=-frnsos1 (s.22)
Since 3o is the unconditional expected value of Zo, given only alive at t =0, it is called the
complete expectation of lfe at birth.s
We recognize that the moments of Zo given above are all unconditional. Conditional mo-
ments, and other conditional measures, are defined in Section 5.3, and the standard actuarial
notation for them is reviewed in Chapter 6.
Exaupr,p 5.3
For the distribution of Example 5.2,find (a) 441 and (b) the median of the distribution.
Sor,urroN
The continuous uniform distribution, defined in Section 2.3.1, is a simple two-parameter dis-
tribution with a constant PDF. fhe parameters of the distribution are the limits;f the interval
5
The significance of the adjective "complete" will become clearer when we consider an alternative measure of
the expectation of life in Sections 5.3.6 and 6.3.4.
68 . CHAPTERFTVE
on the real number axis over which it is defined, and its PDF is the reciprocal of that interval
length. Thus if a generic random variable X is defined over the interval [a,b], then
fx (x) = *, for a I x 1 b, and fx@)= 0 elsewhere.
For the special case of the age-at-failure random variable, a = 0 so b is the length of the in-
terval, as well as the greatest value of I for which fo!)> 0. When the uniform distribution is
used as a survival model, the Greek ar is frequently used for this parameter, so the distribu-
tion is defined by
fol) = *, (s.24)
for 0<t< ot. Tftrc following properties of the uniform distribution easily follow, and should
be verified by the reader:
hQ)=ffi-* (s.27)
o)
Elrof = = (s.28)
lit.7ogat 2
)
or
var (r) = zlro'f - {alrol}' =- (s.2e)
12
The trniform distribution, as a survival model, is not appropriate over a broad range of age, at
least as a model for human survival. It is of historical interes! however, to note that it was the
first continuous probability distribution to be suggested for that puq)ose, n 1724, by Abraham
de Moiwe. As aresulg actuarial literature and exams often refer to the uniform disffibution as
"de Moiwe's law."
The major use of this distribution is over short ranges of time (or age). We will explore this
use of the uniform distribution quite thoroughly in Section 6.5.1.
This very popular one-parameter distribution (see Section 2.3.3) is defined by its SDF to be
So(t1 = ,
(s.30)
"-At
Sunvrv,cr MoDELS . 69
hQ)=ffi=^, (s.32)
a constant. In the actuarial context, where the hazard rate is generally called the force of
mortality, the exponential distribution is referred to as the constantforce distribution.
The exponential distribution, with its properfy of a constant hazardrate, is frequently used in
reliability engineering as a survival model for inanimate objects such as *u"h^io" pu.tr. tit"
the uniform distribution, however, it is not appropriate as a model for human surv'ival over
a
broad range, but might be used over short intervals, such as one year, due to its mathematical
simplicity. This will be explored in Section 6.5.2.
llt = Bct
'
(5.33)
The PDF is given by p, 'So(r), and is clearly not a very convenient mathematical form. A
closed-form expression for the mean of the distribution ,
EfToldoes not exist, but the mean
can be approximated by numerical integration with a large hnite upper limii replacing the
actual upper limit of infinity.
In 1860 Makeham [23] modified the Gompertz distribution by taking the force of mortality
to be
for I > 0, B > 0, c > l,and A> -8. Makeham was suggesting that part of the hazard at any
age is independent of the age itself, so a constant was added to the Gompertz force of mor-
tality.
70 T CHAPTERFTVE
Again it is clear that the PDF for this distribution is not mathematically tractable. As with the
Gompertz distribution, there is no closed-form expression for ElTof, although it can also be
approximated by numerical integration.6
We have briefly explored four distributions here: two (uniform and exponential) which are
mathematically simple, and two (Gompertz and Makeham) which are not. For many illustra-
tions, where we wish to avoid mathematical complexitf, we will use the uniform or the ex-
ponential for illustrative purposes only, not necessarily suggesting that they are applicable in
practice. The exponential has been applied in many situations not involving healthy human
lives, and has been widely used in those situations-
In Section 5.1 we defined a continuous random variable, denoted Zo, which measured the
length of time from age 0 until failure occurs. Now we turn to the case where our entity of in-
terest is known to have survived to age x, where x > 0, and we wish to consider the random
variable for the additional time that the entity might survive beyond age x. We denote this
random variable by T, and note that its domain i. 4 > 0. We define the random variable
{ to be the time-to-failure random variable for an entity known to be alive (i.e., known to
have not yet failed) at age x. We will use the notation (r) to denote the entity known to be
alive at age x.'
If f is the random time-to-failure for an entity alive at age x, it follows that the age-at-
failure will be f more than age x, so we have the relationship Ts= x*7, between our two
basic random variables. This is illustrated in the following figure.
T, +4
x x*7,
FIGURE 5.1
Ratlrer than develop separate distributions for T, for each different value of x, we will simply
calculate probability values fot 4 from the distribution of d. (An exception to this will be
explored in Section 5.4.)
6
A generalization of the Makeham distribution is presented in Exercise 5-10.
7
The time until failure of (x) can also be called the future lifetime of (t), so { is therefore often called the
future lifetime random variable for the entity (.x).
SuRvrvalMoDELS i 71
Recall from basic probability theory that the conditional probability Pr(E lF) is defined to be
Pr(E F\ Pr(EfiF)
I = (s.38)
Pr(F) '
tP' = E0)n
= Pr(Tr> t)
= Pr(To > x+t lTo > x)
(5.3e)
so(x)
E
In the case of human persons under life insurance arrangements, time will normally be measured in years. For
inanimate objects, such as light bulbs or mechanical devices operating under test conditions, time might be meas-
ured in hows.
9
As with the random variable { defined earlier, the notational convention in probability theory would be to
write the SDF symbol as S4(t). To avoid subscripting a subscript, we use the simpler ,Sr(r). This notational prin-
ciple will hold for the CDF, PDF, and HRF as well.
72 I CHAPTERFTVE
te*=Fr(t) =Pr(7,3t)
= Pr(To<x+tlTo>x)
= l- Pr(To > -r+/ | To> x)
So(x+r)
- 1- so(t) ' (5.40)
where we use Equation (5.39) to express Pr(To> x+tlTo >.x) in terms of So(t), the SDF of
the age-at-failure random variable 4. We can easily write Equation (5.40) in terms of
Fo(x). We have
ng(1iO _ rb(.r+r)'-4(r).
"' = r_t,1-4(x)
,e, (s.41)
l-Fo(x)
Since the CDF and SDF are complements of each other, then it follows that tp*=l-,e*
and vice versa.
In the special case of / = 1, the notational convention is to suppress the / part of the symbol.
Thus the probability of (.r) surviving one additional year beyond age x is given by
.f,(t) = 4o,(,)
clt
= -!s,1,1
dt
_ _ d So(o+t)
dt So(r)
- fn$+t) (s.44)
so(") '
since fi (x+ t7 = -SSog+t).
SURVIVAL MoDELS O 73
Considering together Equation (5.44) and Equation (5.10), with x replaced by x+t, we can
express the PDF of { in terms of the distribution of 7o as
In the actuarial context and notation, where the hazard rate is called the force of mortality
and denoted by lt,, we have
age x,lo or the expectedfuture lifetime at age.r, and is denoted by 3,. Recall that the age-at-
failure random variable Zo and the time-to-failure (or future lifetime) random variable {,
arerelatedby 4={,+x, so T*=Ts-x. Thuswehave
Equation (5.47) makes the point that the expected future lifetime of (x) is the excess of the
conditional expected age at failure, E[TolT, > x], over the value of.x itself. It can be calculated
from the conditional PDF of Zo as
o
€x = Il ,. fo@+tlro> x) dt. (s.a8a)
lo
See footnote 5.
74 T CHAPTERFI\E
Next, recall from Equation (5.18) that integration by parts on Equation (5.48b) leads to
o1
Z, = ,fu [l solx+t1dt, (5.48c)
and finally to
r€
[--soE+/)
" = ro
3" dt
srci = Jo 'P'dt'
(5'48d)
by use of Equation (5.39). As we will see in Chapter 6, Equation (5.48d) is a common way
Finally, since the age at failure, given alive at age.r, is the future lifetime, given alive at age
x, plus x itself, then it follows that
and
nlr|lqr"] = Iir''foTlro>x)dy, (5.s1)
provided the two expectations exist. Then the variance of the age at failure and also of the
time to failure is given by
Exrvrpr.r 5.4
Sor-urroN
zoPi6=ffi=ffi# ='82et6'
-- -
tY36 :,so(36+t)
- = .125(64_t\t/2.
s.rgot
o
.rzs[f $4-Dtt2 o, = o
e36 =
?
53.6 DIScnnTE Tn,ffi-TG.F.AILTIRE RANDOM VARIABLES
In this section we present two different, but closely related, discrete random variables. The
first, which we denote by K*, isthe curtate duration atfailure random variable. It is defined
to be the integral part of T*o the future lifetime of (x). The second, which we denote by K|,
is the time interval of failure random variable. Both random variables are illustrated in Fig-
ure 5.2.
K'=k-l
KI =k
2 k-t
Frcunr 5.2
Consider the kth time interval, which we denote by (k-1, /rl to suggest that the precise time
point fr is included in the k'h interval and the precise time point /c - I is included in the
(e-l)t interval. If failure occurs in the kth intewal, including the special case of T,=k,
then Kl = & but K, = k-1, the greatest integer in (. Then it follows that, in general,
K* = Kr+|. (s.53)
In particular, if failure occurs in the first time interval following age x, then we have Kl = 1
for k =1,2,... , and is denoted by r-rl q, rnactuaial notation. The probability function of K, is
= Pr(x+k<4<x+k+Il4>x), (5.55b)
The moments of K, and Kl are all conditional on survival to age x, denoted by 4 > x. The
expected value of K, is given by
It is denoted by e, rnactuarial notation, and is called the curtate expectation of life at age *."
The expectedvalue of Kj is givenbY
and
Var(Kil4 > x) = Var(K*I4 > x), (5.57b)
since K, and ffl differ only by an additive constant. This common variance would be fowrd by
firstfinding either
Next we define the continuous random variable R, to represent the fractional part of the
time interval lived through in the interval of failure for an entity alive at age x. This is illus-
trated in Figure 5.3.
rrNote
the similarity of e, = g1y,l fo >:rl to 2* = EIT,I Zo > xl, defined by Equation (5,47). Recall that the lat-
ter is called th e complete expectation of life at age x whereas the former is called the curtate expectation.
SunrrrverMoDEls o 77
If failure occurs at time r, where 0<r<1, within year k+1, then we have K"=k, Rx=r,
and T, = k+r. In general we have
T, = Kr*Rr.t' (s.sea)
Note that, in the case described in Figure 5.3, the time-interval-of-failure random variable is
Ki = k+1, so the general relationship is
T, = Ki-l+&. (5.seb)
In the life insurance context, a person being issued life insurance coverage at age x is certainly
known to be alive at age x. But additional information is also available about this person as a
result of the underwriting and selection process (see Chapter 4) that people seeking insurance
would undergo. Therefore the probability of surviving on from age x to age x-t t for such a
person would logically be different (ikely greater) than the comparable probability for a per-
son known only to have survived to age x. The probability of surviving from age x to age x + t
for a person underwritten and selectedfor insurance at age.r is denoted by ,p14, to distin-
guish it from ,p,.
The survival probability t p14 ean be represented in survival distribution form. We consider
that the selection event defines time I = 0, and the age at selection, denoted [x], is merely an
identiffing characteristic of the person whose survival distribution we are considering. We
denote such a survival distribution by Strr(l;.x), and denote the random variable for length of
survival by \"1in this case. The identifuing characteristic [x], the age of the person at selec-
tion, is called a concomitant variable. Note that since age .r coresponds to time / = 0, then
rp1,1 is an unconditional probability whereas ,p, is a conditional probability.
The select survival model Sg,1(/;r), in parametric form, would be a function of the two vari-
ables t and x. In actuarial practice, select models are generally presented in tabular form as a
select life table. This is explored further in Section 6.6.
12
We will further discuss the relationship of the random variables 7,, K,, Ki,andR, in Section 6.5.1.
78 . CHAPTERFTVE
Exlupr,n 5.5
Find, in terms of S1r,(t;x), the probability that an entity selected at age x, and known to be
alive at age .r + 10, will fail before age x + 20.
SolurroN
We seek the probability of failure prior to age x + 20, given survival to age .x + 1 0, which is
denoted by roelrpro in standard actuarial notation. It is equal to
When the conditional probabihry ro.ft.r1+ro is multiplied by the probability of the conditioning
event, which is ,S1,1(10;x), the result is the unconditional probability of survival to x+20,
which is ,S1,1(20;x). Thus we have
to41'1+to = 1-toP["]+to
. Si,1(20;-r)
D
S1'1(10;x)
The suwival model presented in this chapter can be represented as a simple, two-state, contin-
uous-time Markov model, where State 1 is continued survival and State 2 is failure (or
d"uth).t' It is clear that the process must begin in State 1 and equally clear that State 2 is an
absorbing state. Therefore there can be only one transition, from State 1 to State 2, over the
entire process. The model is represented by the following diagram.
FIGURE 5.4
The arrow in Figure 5.4 reminds us that transition from State 2back to State I is not possible.
Because the process cannot reenter State I once it has left it, it follows that the event of being in
State 1 at time t * r,
given in State 1 at time /, can occur only if the process never leaves State 1
over that time interval. Then it follows tnat , pil') = ,p{\) , as defined in Section 3.2.
13
In Chapter 6 we will represent the survival model as a discrete-time Markov model.
Sunvrvar,MoDEls | 79
To show the correspondence of the multi-state model notation, defined in Chapter 3, with the
actuarial notation, defined in this chapter, suppose we have a person alive at age x at time r.
The person constitutes the model; since the person is known to be alive, the model is known
to be in State 1. Then the probability of still being in State I at time t * r is given by
The force of transition from State I to State 2 attime s, denoted by fuz!) in Section 3.2.1,
is also d(s), the total force of hansition out of State I since State 2 is the only state to
which the process can move. Note that Ltb) = 0 for all s. In actuarial notation, the force of
transition 4zG) is called the force of faiiure, or the force of mortality, and is denoted simp-
lyby 2(s) orsometimesby 1".
EXAMPLE 5.6
Solve the Kolmogorov differential equation for , p{t) , and translate the result into standard
actuarial notation.
SoLUTIoN
In this case, where j =i =1, the index k in Equation (3.14b) takes on only the value k =2,
since /r * j =1. We have
L rl'll
il) :- o!2
rI'12
. A"r(t+r) -,p{\) . )r(t+r).
O,
d
fr ,n[\' = - ,p[',) 'lr(t+r).
--(/)
1=
p[\)
,
4^,0[\)
dr
= -,2r(t+r),
or
80 T CHAPTERFTVE
Ioo^,0y,, = [i-\tt+ia,
n,o[\,li = ti-4{,*,1a,
^p[lr)
= o-lit{t+ia' , (5.62a)
since op,(!) =1. State I is the surviving state, so np,(l) represents the probability that a pro-
cess (person) alive at age x at time r will be still alive at time t + n, which we call ,p, in
actuarial notation. Also, in actuarial notation, our process (person) is age x+ r attime t + r,
so we would use the notation px+r in place of the multi-state model symbol ),t(t+r). Then
Equation (5.62a) would be written as
Finally, we observe that if the underlying suwival model is exponential, so that the force of
transition function hr!)is a constant, say 2, then we have a homogeneous Markov model,
and the Kolmogorov differential equation is even more easily solved for ,pf!) or , pf) . fn"
details of this are left to the reader as Exercises 5-24 and 5-25.
5.6 Exnncrsns
5.1 The Age-at-Failure Random Variable
5-1 The hazard rate of a survival dishibution is a linear function, h(t)=a+bt, where
a>0 and b>0. Find each of the following:
(a) so(t) (b) ,6(/) (c) The mode of the distribution
SuRvrv.qlMooem r 81
5-2 Let a survival distribution be defined by So(r) = at2 ib, for 0 < / < fr. If the expected
value of Ie is 60, find the median of 4.
5-3 Explain why the hazard rate 1oQ) = e-rt , where r > 0, is not appropriate for a sur-
vival distribution.
5-4 Which of the following could serve as a survival model, for / > 0? If not, why not?
(c) Solt;=s-'2
5-6 Let Xt and X, be independent generic random variables (i.e., not necessarily age-
at-failure random variables). Define the new random variables f =min(Xr, Xr) and
Z =max(Xr,Xr).
(a) Show that Sr(y) is the product ofthe SDF's of X, and Xr.
(b) Show that Fzk) is the product of the CDF's of X, and Xr.
(c) Show that if X1 and X2 both have exponential dishibutions, then tsalso has an
exponential distribution, but Z does not.
5-7 Let the independent generic random variables X, and X, both have exponential
distributions, with parameters ), and, 4, respectively, where 4> )rr. Let Y and Z
be as defined in Exercise 5-6. Given that
^S"(2) =.24
and SzQ)=.86, find the val-
ue of )"1.
(a) Uniform
(b) Exponential
(c) "fo0) =.00125t,
for 0< t <40
82 T CHAPTERFTVB
5-9 Given that )r(t) = k ' tn and AaQ2) = I .26, where 22 is the median age of the distri
bution, find the value of n. (A model with this hazard rate is called a Weibull distri-
bution.)
where pt(t) and p2Q) are polynomials in /. Show that the Gomperz distribution,
given by Equation (5.33), and the Makeham distribution, given by Equation (5.35),
are special cases ofthe generalized distribution.
5-1 1 Given that To has a uniform distribution and that 4e = 42, find Yar(71).
5-13 Given that pr=H, forall t>0, and roPjs=.8l, findthevalue of zopqo.
5-I4 Given that ,So(r)= (l-t)', for O<t<o and r)0, and that ltr=JO and
5-15 Given that )oQ)=(80-t)-ll2, for 0<l<80, find the median of the distribution of
Tzo.
5-16 Giventhat Ss(r) =+, for 0< t<k2 and ft>O, andthat goo
=2.40, findthe
value of 3uo.
5-19 A select survival model is defined by \r,(t;x) = B.rt.c*+'. Show that the select
survival function is of the form
r4r... =
fi,r(r;") .*p[;-4. ("'-r'"'n')]'
^Llnr+lnc. 'J
defined in Section 5.6, with X(0) = 1. Define the random variable {, in terms of the
Markov process X(/).
5-23 Give a formula in multi-state model notation for the expected time spent in State 1,
given a continuous-time Markov survival process for an entity known to be in State
latagexattime /=0.
5-24 Solve Kolmogorov's differential equation for ,pflr) for a homogeneous Markov sur-
vival model, with constant force of transition function &z$)= &(s)= l.