What is queueing theory?
Queueing theory is the study of waiting lines — how they form, how long they get, and how long you spend in them. A queue appears wherever work arrives faster than it can be handled the instant it shows up: shoppers at a checkout, cars at a toll booth, web requests at a server, jobs on a factory floor. From just a few numbers — how often work arrives, how long each item takes to handle, and how many servers are working — queueing theory can predict how a line will behave.
It matters because waiting is waste, and lines behave in ways that defy intuition. Two truths catch most people off guard. First, queues form even when there's spare capacity — irregular arrivals and uneven job sizes guarantee some waiting long before a system is “full.” Second, waiting grows nonlinearly: as a system gets busier the wait creeps up gently, then explodes as it nears 100% utilization — the famous “hockey-stick” curve. Grasping this helps you make better calls with less guesswork: how much capacity to buy, when to add a server, and why one shared line beats several separate ones. This simulator lets you watch those effects rather than just read about them.
Using the simulator
- Two views — use the toggle at the top. Single system animates one queue; Compare queues puts one combined line next to two separate lines, at the same load, so you can see the difference.
- Controls — drag Arrival rate (how often customers arrive) and Service time (how long each takes); in Single view, choose the Number of servers. Everything updates as you drag.
- Play it — Start runs the animation. Pause stops the servers while customers keep arriving (watch the line build) and turns into Resume. Reset clears it.
- Presets — the row of buttons jumps to instructive setups (a calm system, one near overload, and so on) and starts them for you.
- Read it — the panel shows Measured values from the live run beside the exact Theoretical values, and two charts trace average wait and utilization over time. Push the load past what the servers can handle and an UNSTABLE warning appears.
- Change things live — adjust a control mid-run and the line is kept, so you can overload the system, then add a server and watch it recover.