Papers › Accelerated Development of Multicomponent Alloys in Discrete Design Space Using...
Accelerated Development of Multicomponent Alloys in Discrete Design Space Using Bayesian Multi-Objective Optimisation
Osman Mamun, Markus Bause, Bhuiyan Shameem Mahmud Ebna Hai
The archive published only this paper's code-link row. Authors, date and abstract are from arXiv's metadata (CC0), read from the Kaggle arXiv metadata snapshot of 2026-09-12 where its title matched the archive's; the title is the archive's.
Bayesian optimization (BO) protocol based on Active Learning (AL) principles has garnered significant attention due to its ability to optimize black-box objective functions efficiently. This capability is a prerequisite for guiding autonomous and high-throughput materials design and discovery processes. However, its application in materials science, particularly for novel alloy designs with multiple targeted properties, remains limited. This limitation is due to the computational complexity and the lack of reliable and robust acquisition functions for multiobjective optimization. In recent years, expected hypervolume-based geometrical acquisition functions have demonstrated superior performance and speed compared to other multiobjective optimization algorithms, such as Thompson Sampling Efficient Multiobjective Optimization (TSEMO), Pareto Efficient Global Optimization (parEGO), etc. This work compares several state-of-the-art multiobjective BO acquisition functions, i.e., parallel expected hypervolume improvement (qEHVI), noisy parallel expected hypervolume improvement (qNEHVI), parallel Pareto efficient global optimization (parEGO), and parallel noisy Pareto efficient global optimization (qNparEGO) for the multiobjective optimization of physical properties in multi-component alloys. We demonstrate the impressive performance of the qEHVI acquisition function in finding the optimum Pareto front in 1-, 2-, and 3-objective Aluminium alloy optimization problems within a limited evaluation budget and reasonable computational cost. In addition, we discuss the role of different surrogate model optimization methods from a computational cost and efficiency perspective.
Code
Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.
Code Syntology ran Syntology
Not run by Syntology. Nothing on this page verifies that the listed code works.
Results from the paper archive 2025-07-28
No leaderboard rows for this paper in the archive.
Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections