{"id":"arxiv/2607.24643","url":"https://arxiv.org/abs/2607.24643","tags":["math.NA","math.OC"],"score":0,"title":"Anderson acceleration of the proximal point method: the exact adaptive minimax, a spectral phase transition, and optimal safeguarding","domain":"math.NA","source":"arxiv","authors":["Zheng Jia","Yekini Shehu","Yonghong Yao"],"reviews":0,"subject":"math.NA","summary":"\\noindent We study residual-polynomial acceleration of the proximal point method (PPM) for maximal monotone inclusions, with Anderson acceleration (AA) as the prototypical adaptive scheme. We answer three questions exactly. (i)~The minimax complexity over all adaptive methods is precisely $d_0/(K+1)$ per $K$ resolvent evaluations. The upper bound is attained by the averaged-reflection estimator; the matching lower bound uses an explicit skew-adjoint instance with resolvent eigenvalues at the roots of $u^{K+1}=-1$ and $\\csc^2$-distributed masses, on which every degree-$K$ polynomial method…","category":"research","comments":0,"disputed":false,"division":"math","mentions":0,"discussed":0,"published":"2026-07-27","reactions":{}}