Stein's method for the law of large numbers under sublinear expectations (Q2671643)

The author introduces Stein's method on sublinear expectation spaces to prove a convergence rate for \textit{S. Peng}'s law of large numbers [Probab. Uncertain. Quant. Risk 4, Paper No. 4, 8 p. (2019; Zbl 1434.60075)]. The paper is self-contained in the sense that required definitions and results of sublinear expectation theory are included. The main theorem (Theorem 5.1) states that for i.i.d.\ sequences \((X_k)_{k\in\mathbb N}\) on a sublinear expectation space \((\Omega,\mathcal H,\widehat{\mathbb E})\) with \(\bar\mu=\widehat{\mathbb E}[X_1]\), \(\underline{\mu}=-\widehat{\mathbb E}[-X_1]\) and \(\widehat{\mathbb E}[|X_1|^2]<\infty\) it holds that \[ \bigg|\widehat{\mathbb E}\bigg[\varphi\bigg(\frac{X_1+\cdots+X_n}{n}\bigg)\bigg]-\sup_{y\in[\bar\mu,\underline{\mu}]}\varphi(y)\bigg|\leq\frac{C}{\sqrt{n}} \] uniformly for all Lipschitz functions \(\varphi\) with Lipschitz constant at most 1, where \(C\) is a constant depending only on \(\widehat{\mathbb E}[|X_1|^2]\). The author remarks that by similar arguments the error rate \(C_\alpha n^{-\alpha/(1+\alpha)}\) can be shown for \(\alpha\in(0,1]\) and i.i.d.\ random variables with \(\widehat{\mathbb E}[|X_1|^{1+\alpha}]<\infty\). It is worth to mention that a generalization of this result without the assumption of identical distribution has been proven by \textit{M. Hu} et al. [Probab. Uncertain. Quant. Risk 6, No. 3, 261--266 (2021; Zbl 1497.60025)] with a simple probabilistic method which also shows that \(C\) can be chosen as the upper standard deviation.