Answers are grounded in this paper and cite their sources.
High‐entropy alloys (HEAs) offer a vast compositional space with a wide array of potential catalyst formulations. However, exhaustive evaluation of all possible HEA compositions remains computationally and experimentally prohibitive, necessitating the use of optimization algorithms to efficiently identify the most promising candidates. In this work, we explore particle swarm optimization (PSO) as an alternative to the widely applied batch Bayesian optimization (BBO) method. Both algorithms were benchmarked on four established HEA composition‐activity models with respect to their ability to identify the global optimum and to generate datasets suitable for reconstructing the underlying landscape. Under noise‐free conditions, the algorithms demonstrate comparable performance in locating the optimum, while BBO consistently learns the global landscape more effectively. However, in the presence of noise, the Gaussian process surrogate of BBO exhibits overfitting, resulting in excessive exploitation of the identified optimum region and limited exploration of the broader composition space. In contrast, PSO achieves comparable convergence toward the global optimum while maintaining an exploration behavior that is largely unaffected by noise. Consequently, PSO represents a viable and robust alternative to BBO for HEA optimization, particularly in high‐throughput experimental environments where measurements are inherently noisy.
No takes yet. Share an insight, caveat, or question.
Tirmidzi et al. (2026) studied this question.
Synapse has enriched 8 closely related papers on similar clinical questions. Consider them for comparative context: