Attention, denizens of Tinder: Tom Griffiths has a name for your dating life. He calls it "the optimal stopping problem", defined as having a sequence of opportunities to make a decision and trying to work out when you should make that decision.

If that sounds less than romantic, you will forgive him, perhaps, when you learn he is a cognitive scientist, whose life work is figuring out how to apply computer algorithms to everyday human problems.

Griffiths wants us to use computational logic and mathematics to decide when to get off the dating treadmill and marry (the golden formula: look over 37 per cent of the pool and then commit to the next person you meet who is better than anyone you've already met), to decide which restaurant to go to, to decide when to make an offer on a house, to better organise our wardrobes and to schedule our time.

In Sydney to give a TEDx talk called "Everyday algorithms", based on his book with fellow cognitive scientist Brian Christian, Griffiths says humans, unlike computers, "are very good at generalising from small amounts of data".

He gives the example of a small child who is taught the name for a giraffe, say, and is then able to correctly identify a giraffe next time she sees one.

"If we take these things that people can do and put them into computers, we can better understand human reasoning. And if we can understand what computers can do well, it gives us insight into what humans do."

So perhaps it's not that computers are taking our jobs, it's just we need to work with them more and try to see their point of view.

Griffiths says there are ways to incorporate human emotion and spontaneity into computer logic – through randomness, and through game theory, which involves taking into account the unknowability of other people's minds.

Griffiths says that both humans and computers have limited computational resources – so part of best solving a problem is working out how to solve that problem with the minimal "computational cost".

Take the example of deciding on where to dine tonight – or as Griffiths prefers to call it, the "explore/exploit trade-off". Should you exploit the knowledge you have and choose a restaurant you know is decent, or should you go somewhere new on the chance it is better?

"If you are going to be in town for a long time it's more worthwhile to explore. If you're going to leave town soon, there is no point in seeking new information because you won't use it in the future," Griffiths says.

If you're in town for a long time, the optimal strategy is based on something called the "upper confidence bound", which is a mathematical justification for trying new things in the hope there will be something better than you've previously encountered.

Griffiths concedes this strategy could also be used for dating "if your model of romantic life is not a monogamous one".