{"id":"arxiv/2604.25378","url":"https://arxiv.org/abs/2604.25378","tags":["q-fin.PM","cs.CE","math.OC","q-fin.CP"],"score":0,"title":"Yau's Affine-Normal Descent for Large-Scale Unrestricted Higher-Moment Portfolio Optimization","domain":"q-fin.PM","source":"arxiv","authors":["Ya-Juan Wang","Yi-Shuai Niu","Artan Sheshmani","Shing-Tung Yau"],"reviews":0,"subject":"q-fin.PM","summary":"Unrestricted mean-variance-skewness-kurtosis portfolio optimization can capture asymmetry and tail risk, but sample-moment formulations become computationally impractical when the asset universe is large: they produce dense nonconvex quartic objectives with prohibitive coskewness and cokurtosis tensors and anisotropic, ill-conditioned level sets. We develop a structure-exploiting algorithm based on Yau's affine-normal descent that follows affine-normal directions of the current level set while working directly with the return matrix. The method avoids explicit higher-order tensors and…","category":"research","comments":0,"disputed":false,"division":"econ","mentions":0,"discussed":0,"published":"2026-04-28","reactions":{}}