Hello Reader!

Optimization Has a PR Problem!

The moment people hear words like mixed-integer programming or convex optimization, their brain quietly shuts the door.

It sounds abstract. Mathematical. Slightly dangerous.

Yet the funny thing is this: The core ideas behind optimization can be explained with just three surprisingly simple mental pictures.

Today I’ll show you those three metaphors. And at the end I also have a small personal update for the start of the new month.

You need to explain optimization everywhere

If you work in optimization, sooner or later you run into this moment.

• A colleague from another department asks what you’re building.
• A manager asks why the solver ran for six hours.
• A client asks whether the optimization model can really improve their planning.

And suddenly you're not just solving models anymore. You're translating between three different worlds:

• Business decisions
• Mathematical models
• Software systems

I describe my consulting work as me being a translator between those worlds - because I am.

We translate business problems into mathematics… and mathematics into software… and then translate the results back into business language. But translation only works if you speak a language the other person understands.

And that language is almost always metaphors.

The Haystack Conversation

Often, I talk to clients or prospects who are intrigued by optimization. They sense there’s something powerful there, but they alos don’t quite understand and trust what optimization actually does.

If I start explaining branch-and-bound trees or dual bounds, I usually lose them within 20 seconds. So instead I say this:

The naive way would be to pull out every straw one by one until you find the needle. That works… but it would take forever.

Optimization algorithms do something smarter. They divide the haystack in the middle. Then they ask: Which part of the haystack could possibly contain the needle?

Now imagine such algorithm also have an X-ray scanner. Some piles clearly contain no metal at all. You can throw those away immediately. Other piles might contain metal. Those are the ones you divide again and inspect more closely. This continues until eventually one pile becomes so small that the needle is obvious.

That’s essentially what branch-and-bound does.

When clients hear this, something interesting happens. You can almost see the moment where the skepticism fades. And my conclusion that

makes sense to them.

Three Metaphors That Explain Most of Optimization

I challenge you to find a situation in which you are communicating optimization to a layperson in which none of the following metaphors are useful:

1. The Mountain Landscape

Imagine you're hiking through a mountain range in thick fog.

Your goal is to reach the highest peak, but you can only see a few meters around you. So you take a step uphill. Then another. And another.

Eventually you reach a point where every direction goes downhill. Congratulations — you've found a local optimum.

But the real highest mountain might still be somewhere else.

Many optimization algorithms deal with exactly this challenge:
how to avoid getting stuck on the wrong hill.

2. The Traveling Salesman

A classic example from operations research. A salesperson needs to visit many cities exactly once and return home. What is the shortest possible route? With 10 cities, you already have over 150 000 possible tours. And with 20 cities, the number explodes beyond imagination.

This is what we call combinatorial explosion.

Optimization is the art of navigating this enormous space of possibilities efficiently.

3. The Needle in the Haystack

And that brings us back to the haystack as a way of dealing with those large numbers. When the number of possibilities becomes astronomical, brute force simply doesn't work, so we need something more intelligent.

Divide the search space. Discard large parts that cannot contain the optimum. Focus computational effort where the best solutions might actually be hiding.

An Update from Tim

February was a pretty uneventful month for me - just business as usual and resisting the plenty of viruses that go around in Germany this season.

But I've found a new recipe for a chicken-spinach casserole and promptly enjoyed that discovery very much. Also, I'm looking forward to March with its warmer temperatures and the restart of the football season.

I’m curious:

When you explain optimization to someone outside the field, which metaphor do you use?

Hit reply and let me know. I read every response.

Until the next iteration! Tim Varelmann ​Bluebird Optimization​ Complicated Decisions - Simply Automated! Follow me on LinkedIn​

​

P.S.: repeat those three metaphors over and over in your head when brushing your teeth tonight before going to bed - you'll always remember them when you need them next time ;)