skip to content
March, 2023

Metallic glasses (MGs) are created by cooling molten alloys so quickly that crystalline phases do not have time to grow, leaving a solid material with liquid-like structural disorder. This lack of crystalline ordering in MGs gives rise to a variety of desirable properties, including strong corrosion resistance and excellent soft magnetism. Prediction of novel glass-forming alloys is, however, a difficult problem. The size of composition space means even coarse grid-based searches for interesting alloys are infeasible unless heavily constrained, which requires prior knowledge and reduces the possibility of making novel discoveries. Genetic algorithms provide a practical alternative to brute-force searching, by rapidly homing in on fruitful regions and discarding others. In this work, we apply the genetic operators of competition, recombination, and mutation to a population of trial alloy compositions, with the goal of evolving towards candidates with excellent glass-forming ability, as predicted by an ensemble neural-network model. We focus optimization on the maximum casting diameter of a fully glassy rod, the temperature range of the supercooled viscous liquid, and the price-per-kilogramme, to identify commercially viable novel glass-formers. 

Figure caption: Alloy candidates evolve under the action of a genetic algorithm (GA) towards large values of the fully-glassy casting diameter and the temperature range of the supercooled viscous liquid, converging to a Pareto frontier consisting of the best-performing alloys. The Pareto frontier here was identified by the GA in a few minutes, and needed only to consider a fraction of the total search space, presenting a large acceleration relative to brute-force searching.


R. M. Forrest and A. L. Greer, ‘Evolutionary design of machine-learning-predicted bulk metallic glasses’, Digital Discovery, 2 (2023) 202–218

Researcher Profile