# What are Operations Research problems which occur in everyday life?

What are Operations Research problems which occur in your everyday life?

Things that come to mind are for example:

• driving to work: shortest path problem
• packing your backpack for vacation: knapsack problem or bin-packing
• planning for every weather condition: robust optimization

When introducing non-math people to OR, I find it helpful to start with applications from everyday life and then explain that these question also companies have in larger dimensions.

For example packing your backpack is a everyday problem and packing trucks, ships, planes is a generalized problem which is industry-relevant.

• From time to time, I work on scheuling/timetabling.
– T_O
Jun 1 '20 at 20:52
• "driving to work: shortest path problem" with a fastest path variation.
– Mast
Jun 3 '20 at 17:56

Every morning, I update and (warm start) solve my Nonlinear Multivariate Dynamic Stochastic Optimal Control problem to figure out what to do that day.

• Is this a joke :p ? Jun 2 '20 at 6:58
• @Amira Zarglayou No. Don't downvote me for living O.R. All the other answers are small O.R. - that's fine. This is big O.R. Jun 2 '20 at 11:47

Leaving home, running errands, and returning home: TSP, possibly with time windows.

Shopping for clothing, furniture, automobiles: decision making with multiple criteria.

Choosing a check-out lane at a store: queuing theory.

Bagging your own groceries: bin packing.

Washing bird crap off your car: complexity theory (meaning it and trying to prove P = NP are equally futile).

Cutting your lasagna sheets to fit your baking tray with as little waste as possible is a version of the cutting stock problem

I have always thought of getting everything on the grocery list at the store is somewhat of a TSP problem. I think most people (unknowingly) are trying to come up with a solution to the TSP grocery route in their head.

• I have found that solving the TSP for a family shopping typically would take longer than simply walking every aisle and taking what I need as I pass it. Damn complexity! Jun 3 '20 at 12:55
• @Peter-ReinstateMonica, I agree. But I think it comes down to my inferior heuristic. At least it seems it works using a better heuristic (my wife)
– Sune
Jun 4 '20 at 12:45

Hanging up laundry to dry is a form of online bin packing. I usually pick largest items first and pick a "line" with the smallest remaining capacity ("best fit"?)/

This is literally every conversation I have with investors, so here's my \$.02.

Others have described great examples, so I will cover a different aspect:

I have found the use of (everyday) examples makes things more difficult, so I tend to avoid it like the plague.

The reason is that, for a non-math person, once we describe a single application their mind will latch on to that and never let go, especially if they are an authority in something else (e.g. business). If you describe a TSP, then this is suddenly all that OR is about, and that's it.

I have consistently found it excruciatingly difficult to overwrite that idea once it has taken hold. This is very nasty because it can happen in 5 seconds and can consume an entire meeting.

My strategy is to (i) refuse to name applications, or (ii) name at least 20 in a rapid sequence to make the point that the same technology is used for all of these things. In the latter case, I will specifically refuse to go back to any of those 20 examples and describe that in more detail, because I will politely tell my fellow human "that is not what this conversation is about. We can discuss that in another meeting if you want".

Non-math people are often smarter than us, so it's unfair to assume they have to be spoon fed. As long as we don't overwhelm them with jargon, they can keep up just fine, it just takes more skill to bridge the gap seamlessly.

The way I typically do it is to force people to think, so I say things like: "optimisation is a decision-making science. The results of optimisation tell us exactly how to run complex systems. Think about that for a minute".

I then build the conversation based on how they respond. This works because it's suddenly about them, not me.

Once we make a statement, people's mind will seek out the closest thing they know, e.g., "oh so is this like machine learning?". "No".

This can go on for quite a while, and we have to stick to our guns so that people become open to the idea that we are describing a truly new concept, so they can't be "lazy" and reuse their existing knowledge.

It is only after they are able to describe said concept in their own words that I feel safe describing any examples in detail.

• I createt this SE-account specifically to upvote this answer. - It might not be a literal answer to the question, but it is good advice (for all situations where you have to explain something to laypersons) to question the assumption that it is a good idea to use these kind of examples in the first place! Jun 3 '20 at 12:22
• Wow, thanks! :) Jun 3 '20 at 15:10

Woodworking: stock cutting/bin packing.

• 1D for cutting e.g. 1x, 2x dimensional to required length, optimizing for lowest cost - typically priced by volume, but longer stock of equivalent quality can be significantly more expensive
• 2D, guillotine cuts: sheet goods, e.g. cutting plywood to size. Easier to make edge-to-edge cuts, but may waste more material vs. more difficult stopped cuts with less waste
• 3D: cutting and milling smaller 3d parts out of bigger hunks of wood

This is applicable to personal projects (I use it all the time and have written programs specifically to extract reports from SketchUp and optimize my purchases), and is also directly applicable to production-line manufacturing.

Then there is optimizing the order of operations: cutting (tools, order of cuts, switching tools), assembly, prep work, finishing. e.g. can you do some finishing work (e.g. sealing faces of sheet goods) first and start cuts on other stock while it cures?

To me, everything looks like a newsvendor problem.

Obviously:

• How many apples to buy at the grocery store? Newsvendor!

But also:

• How much money to contribute to your healthcare savings account? Newsvendor!
• How much time to allocate for a meeting you're scheduling? Newsvendor!
• How much data to buy in your cell phone plan? Newsvendor!

Pretty much any time you're trying to match an unknown quantity, and the per-unit penalty is different if you're too high vs. too low, it can be modeled as a newsvendor problem.

• Closely related: "The Streetwalker's Dilemma" twitter.com/themarklstone/status/1096458759344394241 Jun 4 '20 at 0:33
• It does not seem that closely related to me, but OK... Jun 4 '20 at 0:36
• It's kind of a sideways version ... kind of. Jun 4 '20 at 0:39
• How? What’s the quantity you’re trying to match, with different penalties for being too high or too low? Jun 4 '20 at 0:42
• Accepting too low an offer yields stockout cost for customer with better offer arriving during busy period vs.holding out for higher offer and wasting "inventory" (getting nothing). Jun 4 '20 at 0:58

If you use public transport to get to and from work you might have several options on from where you board the transit system, where you get off, and what line you take, with a variety of uncertainties regarding the transit system's punctuality and the distance from your home or workplace to the nearest transit stop. There are plenty of optimization problems to think of in this situation - you may want to not walk very far to or from the transit stop, or you want to minimize the risk of missing a connection...