The "optimal stopping problem" faced by house hunting

When it comes to buying or renting a house, this is a headache.

You need to go to various places to inspect the house and learn about the buying information on the market. But whenever you leave a place for viewing, how can you be sure that the next one is better than this one?

In other words, how do you know when to buy it?

In the face of this problem, there may be a very simple solution-37?

If you want to maximize the probability of finding the best home, you should look at all 37 houses on the market. Then when you see the next house, if this house is better than what you saw before, then go ahead.

Or suppose you want to spend a month to find a house, take 37? time, that is, 11 days to establish the standard, and then you can prepare for action.

We know this is feasible because trying to find a good place to stay is a typical "optimal stopping problem." This is a type of problem widely studied by mathematicians and computer scientists.

I am a computational cognitive scientist. I am committed to studying how people's brains work. I will not only study the amazing results of humans, but also the failures.

To this end, I will consider the computational structure of the problems that arise in daily life, and compare the ideal solutions to these problems with the behaviors that we actually do.

At the same time, I also discovered how to use a little knowledge of computational science to make decisions easier for humans.

I have personal motives for this. I have been a serious child since I was a child, and I always try to act in a way that I think is rational. Reasoning about each decision, trying to figure out the best action to take.

Cognitive scientist Tom Griffith as a child

But this method cannot be extended to the various problems you encounter in daily life.

Once, I wanted to break up with my girlfriend because I was thinking about how to find the best compromise between our respective preferences. But in the end it made me exhausted. She pointed out that my solution to this problem was wrong, and later she became my wife.

Regardless of the small decision of which restaurant to go to, it is important enough to decide who to spend the rest of your life with. Human life is always faced with various computing problems that are difficult to solve by effort.

Faced with these problems, consulting an expert is a good choice. Here I am talking about a computer scientist.

When you are looking for advice in life, a computer scientist may not be your preferred conversation partner.

Imagine living like a computer, pursuing unchanging certainty and accuracy. It sounds really boring. But computer science, which studies human decision-making, has revealed such a fact that we reversed it.

"The trade-off between exploration and utilization" in where to eat

When dealing with various problems encountered in people's lives, the way computers solve these problems is very similar to the real human behavior.

Take the example of deciding where to eat. This is a matter of specific computing structure. You have a set of options and you need to choose one of these options. And tomorrow you will face the same decision.

In that case, you encounter the so-called " exploration and exploitation trade-off" in computer science.

You can choose to try new foods, that is, explore, and collect information that you can use in the future; or choose to go to restaurants that have already eaten and taste good, that is, use the information you have collected so far.

When you need to choose between trying new and known high-quality experiences, you are faced with "the trade-off between exploration and utilization." Whether it's listening to music or deciding whom to party with.

Technology companies will also face the same problem. For example, when they need to decide what advertisements should be displayed on the page, should they display new advertisements? Or show the ad that is known to be clicked by the user?

In the past 60 years, computer scientists have made great strides in understanding the trade-off between exploration and utilization. And their research gave some surprising findings.

When you are deciding which restaurant to go to, the first question you should ask yourself is: how long will you stay in the city?

If you are there only for a short stay, then you should enjoy it, that is, "use". There is no point in collecting information, just go directly to the restaurant you like.

But if you will stay for a long time, then take some time to "explore" and try new restaurants, because the information you collect can optimize your choices in the future. The more valuable information increases, the more opportunities you have to use it in the future.

This principle can al