BIRTH AND DEATH PROCESS:
The birth and death process is a special case of continuous time Markov process, where the
states represent a current size of a population and the transitions are limited to birth and
death. When a birth occurs, the process goes from state n to state n + 1. Similarly, when
death occurs, the process goes from state n to state n − 1. It is assumed that the birth and
death events are independent of each other. The birth-and-death process is characterized
by the birth rate λn and death rate µn which vary according to state n of the system.
A random process {N (t); t ≥ 0} is said to be a birth-death process if it satisfies the following
assumptions:
During time interval [t, t + δt] where δt is a sufficiently small time interval
(i) Pr[N (t + δt) − N (t) = 1|N (t) = n] = λn δt + ◦(δt)
(ii) Pr[N (t + δt) − N (t) = −1|N (t) = n] = µn δt + ◦(δt)
(iii) Pr[N (t + δt) − N (t) = 0|N (t) = n] = 1 − λn δt − µn δt + ◦(δt)
(iv) Pr[X(t + δt) − X(t) ≥ 2 or ≤ −2|X(t) = n] = ◦(δt)
where
◦(δt)
lim =0
δt→0 δt
We use these assumptions to define system of equations that are useful in deriving prop-
erties of population size at given time t.
1
Probability distribution of N(t):
Let Pr[N (t) = n] = Pn (t) and Pr[N (t + δt) = n] = Pn (t + δt).
If the population size at time t + δt is n, then the possible values of population size at time
t are either n − 1,n or n + 1; n = 1, 2, ...
Therefore
Pn (t + δt) = Pr[N (t + δt) = n]
= Pr[N (t + δt)) = n, X(t) = n] + Pr[N (t + δt)) = n, N (t) = n − 1] +
Pr[N (t + δt)) = n, N (t) = n + 1]
= Pr[N (t + δt) − N (t) = 0|N (t) = n]Pr[N (t) = n] +
Pr[N (t + δt) − N (t) = 1|N (t) = n − 1]Pr[N (t) = n − 1] +
Pr[N (t + δt) − N (t) = −1|N (t) = n + 1]Pr[N (t) = n + 1]
= (1 − λn δt − µn δt + ◦(δt))Pn (t) + (λn−1 δt + ◦(δt))Pn−1 (t) +
(µn+1 δt + ◦(δt))Pn+1 (t)
Pn (t + δt) − Pn (t) = −(λn δt + ◦(δt))Pn (t) + (λn−1 δt + ◦(δt))Pn−1 (t) +
(µn+1 δt + ◦(δt))Pn+1 (t)
Dividing through by δt
Pn (t + δt) − Pn (t) δt δt ◦(δt) δt ◦(δt)
= −λn − µn + Pn (t) + λn−1 + Pn−1 (t) +
δt δt δt δt δt δt
δt ◦(δt)
µn+1 + Pn+1 (t)
δt δt
Pn (t + δt) − Pn (t) ∂Pn (t)
lim = = Pn0 (t)
δt→0 δt
∂t
δt ◦(δt)
lim λn−1 + Pn−1 (t) = λn−1 Pn−1 (t)
δt→0 δt δt
δt ◦(δt)
lim µn+1 + Pn+1 (t) = µn+1 Pn+1 (t)
δt→0 δt δt
δt δt ◦(δt)
lim −λn − µn + Pn (t) = −(λn + µn )Pn (t)
δt→0 δt δt δt
Combining all these
Pn0 (t) = − (λn + µn ) Pn (t) + λn−1 Pn−1 (t) + µn+1 Pn+1 (t) n = 1, 2, ...
2
Now what if the population size at time t + δt is zero, then the possible values of pop-
ulation size at timt tare zero and one.
P0 (t + δt) = Pr[N (t + δt) = 0]
= Pr[N (t + δt)) = 0, N(t) = 0] + Pr[N (t + δt)) = 0, N(t) = 1]
= Pr[N (t + δt) − N (t) = 0|N (t) = 0]Pr[N (t) = 0] +
Pr[N (t + δt) − N (t) = −1|N (t) = 1]Pr[N (t) = 1]
= (1 − λ0 δt + ◦(δt))P0(t) + (µ1 δt + ◦(δt))P1 (t)
P0 (t + δt) − P0 (t) = −(λ0 δt + ◦(δt))P0(t) + (µ1 δt + ◦(δt))P1(t)
Dividing through by δt
P0 (t + δt) − P0 (t) δt ◦(δt)
= −λ0 + P0 (t) +
δt δt δt
δt ◦(δt)
µ1 + P1 (t)
δt δt
P0 (t + δt) − P0 (t) ∂P0 (t)
lim = = P00 (t)
δt→0 δt ∂t
δt ◦(δt)
lim −λ0 + P0 (t) = −λ0 P0 (t)
δt→0 δt δt
δt ◦(δt)
lim µ1 + P1 (t) = µ1 P1 (t)
δt→0 δt δt
Combining all these
P00 (t) = −λ0 P0 (t) + µ1 P1 (t) n = 0
The two equations
P00 (t) = −λ0 P0 (t) + µ1 P1 (t) n = 0
Pn0 (t) = − (λn + µn ) Pn (t) + λn−1 Pn−1 (t) + µn+1 Pn+1 (t) n = 1, 2, ...
are referred to as basic difference-differential equations for birth and death process.
For specified initial conditions for the process, a solution for Pn (t) can be obtained but it
is rarely obtained in an analytical form.
3
Steady state probabilities
Since explicit solutions of Pn (t) are not easily obtained, the steady state solution is of special
interest in an effort to solve birth and death processes. The steady state solution assumes
that after a sufficient large time has elapsed, the process attains statistical equilibrium. In
a steady state, all state porbabilities Pn (t) be constant and hence the derivatives Pn0 (t) are
all equal to zero. Therefore under steady state assumptions,
0 = − (λn + µn ) Pn (t) + λn−1 Pn−1 (t) + µn+1 Pn+1 (t) n = 1, 2, ...
0 = −λ0 P0 (t) + µ1 P1 (t) n = 0
Using iterative method:
For n=0
0 = −λ0 P0 (t) + µ1 P1 (t)
λ0 P0 (t) = µ1 P1 (t)
λ0
P1 (t) = P0 (t)
µ1
For n=1
0 = − (λ1 + µ1 ) P1 (t) + λ0 P0 (t) + µ2 P2 (t)
(λ1 + µ1 ) P1 (t) = λ0 P0 (t) + µ2 P2 (t)
(λ1 + µ1 ) P1 (t) = µ1 P1 (t) + µ2 P2 (t)
λ1 P1 (t) + µ1 P1 (t) − µ1 P1 (t) = µ2 P2 (t)
λ1 P1 (t) = µ2 P2 (t)
λ1
P2 (t) = P1 (t)
µ2
λ1 λ0
P2 (t) = P0 (t)
µ2 µ1
λ1 λ0
P2 (t) = P0 (t)
µ2 µ1
Proceeding in this manner , we obtain
λn−1 λn−2 .....λ0
Pn (t) = P0 (t)
µn µn−1 ....µ1
