For algorithms, memory is a much stronger resource

This classic result was a way to convert any algorithm with a certain time budget in a new algorithm with a slightly smaller space budget. Williams saw that sponge gravel simulation will make the use of the new algorithm area much smaller – equal to the square root of the budget of the original algorithm time. The new space -saving algorithm would be much slower, and therefore the simulation was not likely to have practical applications. But from the theoretical point of view, it was not less than the revolutionary.

For 50 years, the researchers assumed that it was impossible to improve the Hopecroft, Paul and Valiant global simulation. Williams idea – if she succeeds – will not overcome their record – it will destroy it.

“I thought about it, and I was like,” well, this simply cannot be true. “Williams said.” He put it aside and did not return to that fateful day in July, when he tried to find the defect in the argument and failed.

At the end of February, finally Williams Put the final paper on the Internet. Cook and Mertz were surprised like anyone else. “I had to go for a long time before doing anything else,” said Mertz.

Valiant got an infiltration inspection to improve Williams on his decades -long result. For years, he taught at Harvard University, on the road from Williams office at the Massachusetts Institute of Technology. They met before, but they did not know that they lived in the same neighborhood until they collided with each other on the bus on a snowy day in February, a few weeks before the result. Williams described his evidence of amazing courage and promised to send his paper.

“I was very impressed,” said Valiant. “If you get any mathematical result, it is the best thing in 50 years, you must do something right.”

Pspace: final limits

Through his new simulations, Williams have proven a positive result about the mathematical strength of the area: algorithms that use relatively small space can solve all problems that require more time. After that, using a few lines of mathematics, turned on it and has proven a negative result about the strength of the calculation of time: some problems can not be solved at least unless you use more time than space. This second, the narrower result is in line with what the researchers expected. The strange part is how Williams has reached there, by proving the result that applies to all algorithms, regardless of the problems they reside.

“I still have it difficult to believe,” Williams said. “It seems very good to be true.”

Quality formulation, the second result of Williams may look like a long -tested solution for the PSPace problem. The difference is the issue of size. P and PSPace are very wide complications, while Williams results are working more accurate. He established a quantitative gap between space strength and time force, and to prove that PSPace is larger than P, researchers will have to make this gap much broader.

This is a hard challenge, similar to getting rid of a sidewalk with a goal until it becomes wide like Grand Canyon. But it may be possible to reach there using a modified version of the Williams simulation that repeats the main step several times, providing a small space at a time. It is like a way to frequently increase the length of the scattered – make it large enough, and you can open anything. This repeated improvement does not work with the current release of algorithm, but researchers do not know whether this is a basic restriction.

“The bottleneck may be the final bottle, or the bottleneck may be for 50 years,” Valenant said. “Or it might be something that someone can solve next week.”

If the problem is solved next week, Williams will kick himself. Before writing the paper, he spent months trying and failed to extend his results. But even if this extension is not possible, Williams is sure that more exploring the space must lead to an interesting place – it may be presented in a completely different problem.

“I can never prove the things I want to prove specifically,” he said. “But often, the thing that has proven is better than what I wanted.”

Editor Note: Scott Aaronson is a member of Quanta’s magazine Consultative Council.

---

The original story Recal it with permission from Quanta magazineAn independent editorial publication for Simonz Foundation Its mission is to enhance the general understanding of science by covering research developments and trends in mathematics, physical sciences and life.