Operational research in everyday life – Model thinking
Welcome to the first in a four-part series on how operational research and model thinking shape everyday decisions, from the lines we wait in to the routes we drive. New installments will follow at three-month intervals, unpacking how simple, well-built models help us see complexity clearly—and act with confidence.
Who’s behind this series?
The perspective here is informed by a semi-retired Massachusetts Institute of Technology (MIT) professor who has spent a career moving across five departments in engineering and in the School of Architecture and Planning—now based at the Institute for Data, Systems, and Society. That breadth matters: it’s where model thinking thrives. In their words, operations research (called operational research in the UK) is “the world’s most important invisible profession.” Born in World War II and, in spirit, a kind of physics for human-made systems, it studies how real operations—networks, inventories, schedules, services—behave, and how to improve them using data and models.
What is “model thinking”?
Model thinking has two complementary meanings:
- Using mathematical or conceptual models to clarify how a system works and to make better decisions.
- Thinking as a “model” to emulate—clear, disciplined, transparent, and testable.
In practice, the best decisions blend both: rigorous use of models and exemplary reasoning about assumptions, limits, and trade-offs.
Three everyday applications
1) Queues: why waiting lines behave the way they do
From coffee shops to call centers and emergency rooms, queues are everywhere. Queueing theory—a core pillar of operational research—was pioneered by Danish telephone engineer A. K. Erlang in 1919 to analyze how many lines a switchboard needed. His formulas still underpin performance predictions today by relating a few vital parameters:
- Arrival rate: how quickly customers show up
- Service time: how long each customer takes
- Number of servers: how many parallel “lanes” handle demand
Change any one of these, and waiting times and congestion shift in predictable ways. That’s powerful: with a small model, you can test “what if” scenarios before investing in more staff or capacity. Modern uses span telecoms, hospitals, cloud computing, and transportation.
2) Inventory: having enough, but not too much
Inventory theory asks: when should you reorder, and by how much, to keep service high without tying up cash or space? The heart of it is a model of stock dynamics under uncertain demand and variable lead times. Common levers include:
- Reorder points: a threshold that triggers a new order
- Safety stock: a buffer against demand spikes or delays
- Order quantities: balancing holding costs against ordering costs
With a basic model, you can set rules that minimize stockouts while keeping carrying costs in check—whether you manage a warehouse, a pharmacy, or even a home pantry.
3) Routing: the math behind your satnav
Tap “start” on your car’s dashboard and, in milliseconds, an algorithm weighs countless paths to suggest the one that minimizes travel time. A foundational tool here is Dijkstra’s algorithm, discovered by Edsger W. Dijkstra in 1956 and published a few years later. It efficiently finds the shortest path on a network—now augmented by real-time traffic data, turn penalties, and other constraints in modern navigation systems.
Why it matters: from answers to understanding
Model thinking pairs naturally with discovery learning: instead of searching for a quick answer, you build a small model of the process and test it yourself. That effort cements understanding. You learn where a model breaks, what assumptions matter, and how to adapt it. As the well-known adage—often linked to Benjamin Franklin—suggests: tell me, and I may forget; involve me, and I learn.
Try it yourself
- Queues: Time five arrivals and five services at your local cafe. Estimate the arrival rate and average service time. Consider how adding a barista or changing workflow would alter wait times.
- Inventory: Track your household consumption of a staple (e.g., coffee beans) for two weeks. Set a reorder point so you never run out, factoring delivery lead time and a small buffer.
- Routing: Compare your navigation app’s “fastest” route with an alternative at a different time of day. Note how traffic data changes edge weights and reshapes the recommended path.
Good modeling habits
- State objectives clearly: minimize waiting, avoid stockouts, reduce travel time.
- Write down assumptions: arrival patterns, demand variability, data sources.
- Start simple: build a small, testable model before adding complexity.
- Validate: compare model predictions with observed outcomes; iterate.
- Explain trade-offs: faster service vs. higher cost; lower stockouts vs. more inventory.
The big picture
Operational research brings disciplined clarity to messy, real-world systems. It’s not about mystique or giant datasets; it’s about framing questions carefully, using the right simple model, and improving decisions step by step. In the coming entries, we’ll continue exploring how this “physics of operations” helps individuals, teams, and institutions perform better—with tools you can apply the same day you learn them.
