{"id":"arxiv/2604.02917","url":"https://arxiv.org/abs/2604.02917","tags":["math.OC","cs.CE","cs.DC","cs.LG","math.NA"],"score":0,"title":"Scalable Mean-Variance Portfolio Optimization via Subspace Embeddings and GPU-Friendly Nesterov-Accelerated Projected Gradient","domain":"math.OC","source":"arxiv","authors":["Yi-Shuai Niu","Yajuan Wang"],"reviews":0,"subject":"math.OC","summary":"We develop a sketch-based factor reduction and a Nesterov-accelerated projected gradient algorithm (NPGA) with GPU acceleration, yielding a doubly accelerated solver for large-scale constrained mean-variance portfolio optimization. Starting from the sample covariance factor $L$, the method combines randomized subspace embedding, spectral truncation, and ridge stabilization to construct an effective factor $L_{eff}$. It then solves the resulting constrained problem with a structured projection computed by scalar dual search and GPU-friendly matrix-vector kernels, yielding one computational…","category":"research","comments":0,"disputed":false,"division":"math","mentions":0,"discussed":0,"published":"2026-04-03","reactions":{}}