Scientists Prove Perfect Online Matching Is Impossible, Even for AI
This could affect how companies match you to jobs, dates, or ads.
Imagine you run a ride-sharing app. Drivers are available, and ride requests pop up randomly. You want to match each request to the best driver, but you don't know future requests. This is called online stochastic matching. For years, researchers wondered: can a computer always find the best matching strategy? A new study says no—it's computationally hard, even for powerful computers.
The researchers proved that for random-order arrivals, finding the optimal online strategy is so complex that no efficient algorithm can solve it exactly, unless a major mathematical breakthrough occurs (P=NP). This means companies can't guarantee perfect matches; they must use approximations. For you, that might mean slightly longer wait times for a ride or less relevant ads. But it also means companies can't manipulate the system to always give you the best outcome—they're limited by math.
In practical terms, this research sets a boundary on what's possible. It tells engineers not to waste time seeking perfect solutions and instead focus on good-enough ones. For the rest of us, it's a reminder that some problems are inherently hard, and technology has limits. The next time an app matches you with something, know that behind the scenes, it's doing its best under difficult constraints.
- Finding the best way to match things online is computationally impossible for computers to do perfectly.
- Companies must use approximations, which can lead to slightly worse matches and higher costs.
- This settles a long-standing question but doesn't change everyday technology much—it just sets limits.
Why It Matters
Companies can't guarantee perfect matches, so you might get slightly worse recommendations or service.