Birth-Death Processes — Mathematics Knowledge System

Overview

A birth-death process is a continuous-time Markov chain on the non-negative integers {0, 1, 2, ...} in which transitions can only move to adjacent states: from state n, the process can jump to n+1 (a 'birth') at rate lambda_n or to n-1 (a 'death') at rate mu_n. This structure makes the stationary distribution computable via detailed balance. Birth-death processes model population dynamics, queueing systems (M/M/1), chemical reactions, and epidemic spread.

Intuition

Imagine a population that can only grow or shrink by one individual at a time. While in state n, births arrive at rate lambda_n and deaths at rate mu_n. Because only nearest-neighbor jumps are allowed, the chain is tridiagonal and detailed balance gives explicit formulas for the stationary distribution. The M/M/1 queue is the canonical example: customers arrive at rate lambda and are served at rate mu.

Formal Definition

Definition

A birth-death process is a CTMC on S = {0, 1, 2, ...} with generator Q given by: q_{n,n+1} = lambda_n (birth rate from state n), q_{n,n-1} = mu_n (death rate from state n, mu_0 = 0), and q_{nn} = -(lambda_n + mu_n). The detailed balance equations pi_n * lambda_n = pi_{n+1} * mu_{n+1} have an explicit solution.

q_{n,n+1} = \lambda_n \ge 0, \quad q_{n,n-1} = \mu_n \ge 0 \;(\mu_0 = 0)

Transition rates

\pi_n \lambda_n = \pi_{n+1} \mu_{n+1} \quad \text{(detailed balance)}

Detailed balance

\pi_n = \pi_0 \prod_{k=0}^{n-1} \frac{\lambda_k}{\mu_{k+1}}

Stationary distribution (product form)

\pi_0 = \left(1 + \sum_{n=1}^\infty \prod_{k=0}^{n-1} \frac{\lambda_k}{\mu_{k+1}}\right)^{-1}

Normalization constant

Notation

Notation	Meaning
$\lambda_n$	Birth rate in state n (rate of jumping from n to n+1)
$\mu_n$	Death rate in state n (rate of jumping from n to n-1), with mu_0 = 0
$\rho$	Traffic intensity for M/M/1: rho = lambda/mu

Theorems

Theorem 1: Theorem 1 (Stationary Distribution via Detailed Balance)

\text{A birth-death process has stationary distribution } \pi_n = \pi_0 \prod_{k=0}^{n-1} \frac{\lambda_k}{\mu_{k+1}} \text{ provided the sum } \sum_{n=0}^\infty \prod_{k=0}^{n-1} \frac{\lambda_k}{\mu_{k+1}} < \infty.

Theorem 2: Theorem 2 (M/M/1 Queue)

\text{For the M/M/1 queue with constant rates } \lambda_n = \lambda, \mu_n = \mu \text{ (n} \ge 1\text{), if } \rho = \lambda/\mu < 1, \text{ the stationary distribution is geometric: } \pi_n = (1-\rho)\rho^n, \;n \ge 0. \text{ The mean queue length is } \mathbb{E}[N] = \rho/(1-\rho).

Worked Examples

1
Compute the traffic intensity rho = lambda/mu = 3/5 = 0.6 < 1, so the queue is stable.
$\rho = \frac{\lambda}{\mu} = \frac{3}{5} = 0.6$
2
The stationary distribution is geometric: pi_n = (1-rho)*rho^n.
$\pi_n = (1 - 0.6)(0.6)^n = 0.4 \cdot (0.6)^n, \quad n = 0, 1, 2, \ldots$
3
Mean number of customers: E[N] = rho/(1-rho) = 0.6/0.4 = 1.5.
$\mathbb{E}[N] = \frac{\rho}{1-\rho} = \frac{0.6}{0.4} = 1.5$

✓ Answer

pi_n = 0.4 * 0.6^n, E[N] = 1.5 customers.

Practice Problems

Mediumfree response

For an M/M/1 queue with rho = lambda/mu < 1, verify that the geometric distribution pi_n = (1-rho)*rho^n satisfies detailed balance.

Mediumfree response

An M/M/1/K queue has finite capacity K (the system holds at most K customers). Arrival rate lambda, service rate mu. Find the stationary distribution.

Common Mistakes

Common Mistake

Forgetting that mu_0 = 0 (no death from state 0)

State 0 is a boundary: you can only leave by a birth. Setting mu_0 = 0 prevents the process from going below 0.

Common Mistake

Applying M/M/1 results when rho >= 1

The M/M/1 queue is stable only when rho = lambda/mu < 1. If rho >= 1, no proper stationary distribution exists and the queue grows without bound.

Quiz

Summary

Birth-death processes are CTMCs on {0, 1, 2, ...} with only nearest-neighbor jumps.
Detailed balance pi_n * lambda_n = pi_{n+1} * mu_{n+1} gives explicit product-form stationary distributions.
M/M/1 queue: stationary distribution is geometric pi_n = (1-rho)*rho^n when rho = lambda/mu < 1.
Mean M/M/1 queue length: E[N] = rho/(1-rho).
Linear death rates mu_n = n*mu with constant births yield a Poisson stationary distribution.

References

BookNorris, J. R. -- Markov Chains, Chapter 3
BookKarlin, S. and Taylor, H. -- A First Course in Stochastic Processes, Chapter 4
WebsiteWikipedia -- Birth-death process