Bayesian optimisation with transfer learning for NASICON-type solid electrolytes for all-solid-state Li-metal batteries.

NASICON-type LiZr2(PO4)3 (LZP) has attracted significant attention as a solid oxide electrolyte for all-solid-state Li-ion or Li-metal batteries owing to its high Li-ion conductivity, usability in all-solid-state batteries, and electrochemical stability against Li metal. In this study, we aim to improve the Li-ion conductivity of Li-rich NASICON-type LZPs doped with CaO and SiO2, i.e., Li1+x+2y Ca y Zr2-y Si x P3-x O12(0 ≤ x ≤ 0.3, 0 ≤ y ≤ 0.3) (LCZSP). Herein, a total of 49 compositions were synthesised, and their crystal structures, relative densities, and Li-ion conductivities were characterised experimentally. We confirmed the improvement in Li-ion conductivity by simultaneous replacement of Zr and P sites with Ca and Si ions, respectively. However, the intuition-derived determination of the composition exhibiting the highest Li-ion conductivity is technically difficult because the compositional dependence of the relative density and the crystalline phase of the sample is very complex. Bayesian optimisation (BO) was performed to efficiently discover the optimal composition that exhibited the highest Li-ion conductivity among the samples evaluated experimentally. We also optimised the composition of the LCZSP using multi-task Gaussian process regression after transferring prior knowledge of 47 compositions of Li1+x+2y Y x Ca y Zr2-x-y P3O12 (0 ≤ x ≤ 0.376, 0 ≤ y ≤ 0.376) (LYCZP), i.e., BO with transfer learning. The present study successfully demonstrated that BO with transfer learning can search for optimal compositions two times as rapid as the conventional BO approach. This approach can be widely applicable for the optimisation of various functional materials as well as ionic conductors.

Bayesian Optimization for Cascade-Type Multistage Processes.

Complex processes in science and engineering are often formulated as multistage decision-making problems. In this letter, we consider a cascade process, a type of multistage decision-making process. This is a multistage process in which the output of one stage is used as an input for the subsequent stage. When the cost of each stage is expensive, it is difficult to search for the optimal controllable parameters for each stage exhaustively. To address this problem, we formulate the optimization of the cascade process as an extension of the Bayesian optimization framework and propose two types of acquisition functions based on credible intervals and expected improvement. We investigate the theoretical properties of the proposed acquisition functions and demonstrate their effectiveness through numerical experiments. In addition, we consider suspension setting, an extension in which we are allowed to suspend the cascade process at the middle of the multistage decision-making process that often arises in practical problems. We apply the proposed method in a test problem involving a solar cell simulator, the motivation for this study.