| Abstract Scope |
The need for rapid materials development to address socio-technological challenges is critical. Traditional methods, such as ICME and high-throughput combinatorial frameworks, are inefficient and resource-intensive. BIRDSHOT (Batch-wise Improvement in Reduced Materials Design Space using a Holistic Optimization Technique) addresses these limitations as it: (i) combines advanced simulations, physics-based and ML models to efficiently identify the feasible regions amenable to optimization; (ii) exploits correlations to fuse simulations; and (iii) uses batch Bayesian Optimization (BO) to make globally optimal parallel iterative decisions on where to explore/exploit the HEA space, leveraging existing models and data. Here, we show how BIRDSHOT successfully guided the discovery of optimal CoCrFeNiVAl FCC HEAs, achieving a 200 to 2,000-fold acceleration in alloy discovery. This was accomplished by exploring only 0.16% of the design space, demonstrating the framework’s capability to optimize multiple objectives and constraints simultaneously. |