Free Queueing Theory Calculator

Queueing Theory Explained: From Erlang to the M/M/s Model

Queueing theory is the mathematical discipline that analyzes waiting lines (queues) using probabilistic models. Originating from the work of Danish engineer Agner Krarup Erlang in the early 1900s—who studied telephone network congestion—this framework now underpins capacity planning in countless fields, from retail and healthcare to telecommunications and computing. A dedicated queue analysis tool such as the Queueing Theory Calculator makes these concepts practical by automatically computing key performance metrics based on a few input parameters.

What Is Queueing Theory?

Queueing theory predicts the behavior of waiting lines over the long run, focusing on the steady state where statistical averages stabilize. Core measures include:

• the average number of customers in the queue or in the entire system,
• the average waiting and service times,
• the utilization rate of servers, and
• the probability of encountering a queue of a given length.

Because arrivals and service completions involve randomness, the theory relies on stochastic processes—most commonly the Poisson process for arrivals and exponential distributions for service times.

Key Elements of a Queue and Kendall Notation

Every queue can be described by four basic components:

• Arrival process (how customers enter the line),
• Service process (how long customers occupy a server),
• Number of servers (resources that process customers in parallel),
• Queue discipline (the order in which customers receive service; FIFO is the default).

To classify different queue types, the Kendall notation uses the form $A/B/c$. Letters $A$ and $B$ denote the probability distributions of inter‑arrival times and service times, while $c$ is the number of servers. The most common symbols are:

• M – Markovian (exponential) distribution,
• E – Erlang distribution,
• G – General (arbitrary) distribution with known mean and variance.

Extended notation $A/B/c/K/N/D$ can include queue capacity, population size, and discipline. The simplest and most studied model is the M/M/1 queue: exponential arrivals and service with one server.

The M/M/1 Queue: A Single‑Server Model

The M/M/1 queue assumes:

• arrivals follow a Poisson process with rate $\lambda$ (average arrivals per time unit),
• service times are exponentially distributed with rate $\mu$ (average completions per time unit when the server is busy),
• a single server with infinite buffer capacity,
• FIFO discipline.

For the system to be stable, the traffic intensity must satisfy $\rho = \lambda / \mu < 1$. If $\rho \ge 1$, the queue grows without bound and never reaches steady state.

The key queuing theory formulas for M/M/1 are derived from a simple Markov chain and Little’s law:

• Server utilization: $\rho = \dfrac{\lambda}{\mu}$
• Average number in system: $L = \dfrac{\rho}{1 - \rho}$
• Average time in system: $W = \dfrac{1}{\mu - \lambda}$
• Average number in queue: $L_q = \dfrac{\rho^2}{1 - \rho}$
• Average time in queue: $W_q = \dfrac{\rho}{\mu - \lambda}$
• Probability of empty system: $p_0 = 1 - \rho$
• Probability of $n$ customers: $p_n = \rho^n (1 - \rho)$

These relationships allow you to quickly evaluate a waiting line’s performance using nothing more than $\lambda$ and $\mu$.

The M/M/s Queue: Multiple Servers

When more than one server is available, the model becomes M/M/s. Each server works independently at rate $\mu$, giving a combined service capacity of $s\mu$. Traffic intensity is redefined as $\rho = \lambda / (s\mu)$; again, $\rho < 1$ is required for stability.

The steady‑state probability of zero customers in the system takes a more complex form:

$$p_0 = \left[ \sum_{k=0}^{s-1} \frac{(s\rho)^k}{k!} + \frac{(s\rho)^s}{s!} \cdot \frac{1}{1 - \rho} \right]^{-1}$$

From $p_0$ you can compute the average queue length:

$$L_q = \frac{ (s\rho)^s \rho }{ s! \, (1-\rho)^2 } \cdot p_0$$

The average number in the system is $L = L_q + s\rho$, and waiting times follow from Little’s law:

$$W_q = \frac{L_q}{\lambda}, \qquad W = \frac{L}{\lambda}$$

An important additional metric is the probability that an arriving customer finds all servers busy and must wait – the Erlang C formula:

$$C(s, \alpha) = \frac{ \frac{(s\rho)^s}{s!} \cdot \frac{1}{1-\rho} }{ \sum_{k=0}^{s-1} \frac{(s\rho)^k}{k!} + \frac{(s\rho)^s}{s!} \cdot \frac{1}{1-\rho} }$$

where $\alpha = \lambda / \mu$.

How to Use the Queueing Theory Calculator

The Queueing Theory Calculator acts as an M/M/1 queue calculator and an M/M/s queue calculator in one interface. Simply choose the model, then input the known parameters:

• arrival rate $\lambda$,
• service rate $\mu$,
• for M/M/s: number of servers $s$.

The tool instantly returns utilization $\rho$, average system length $L$, average queue length $L_q$, average waiting times $W$ and $W_q$, and the probability that the system is empty ($p_0$). For M/M/s it also provides the Erlang C probability.

This waiting line calculator is ideal for performing “what‑if” analyses—for example, seeing how adding one more server reduces waiting times, or how a change in arrival rate affects queue length. It turns abstract formulas into decision‑ready data, making it an essential queue analysis tool for operations managers and students alike.

Applications Beyond Human Waiting Lines

Queueing theory extends far beyond supermarket queues. It models:

• Telecommunications: packet switching, call‑center staffing,
• Computing: process scheduling, memory management,
• Manufacturing: production line balancing,
• Healthcare: patient flow, appointment scheduling,
• Transportation: airport runway use, traffic flow.

By mastering the fundamentals and applying them with a reliable calculator, you can optimize service efficiency, reduce customer wait times, and make informed resource‑allocation decisions.