Improving Advanced Materials Design With a Game Algorithm

If there exists a truly complicated process, it is the design of advanced materials. This is because the potential combinations for precisely placing atoms within just one structure are seemingly endless. But now, thanks to a newly developed game algorithm, the process has become much easier.
To improve the process of designing advanced materials, Japanese researchers have developed a method called Materials Design using Tree Search (MDTS). To identify the best atomic positions in a structure, this method employs an algorithm called the Monte Carlo tree search, which is essentially a game algorithm.
If you’re wondering what a game algorithm has to do with materials design, The Monte Carlo tree search has been successfully used by computer games developers to determine moves in games that bring the best possible outcomes.
In materials design, the algorithm is used to identify the best way to design silicon-germanium alloy structures. And compared to another commonly used algorithm for this purpose, the MDTS was comparable or better regarding total computational time. To top it off, the new method can learn from data and become better in time.

Source:
Phys.org (https://phys.org/news/2017-08-game-algorithm-materials.html)

Advertisement