Searcharxiv⌕ Search

arXiv · 2611.13905

An Accelerated Dynamic Stochastic Primal-Dual Method for Multi-stage Stochastic Optimization

Abstract

We consider multi-stage stochastic optimization problems with smooth convex stage costs and linear linking constraints. Dynamic stochastic approximation (DSA), a stochastic first-order method, uses a recursive structure whose memory requirement grows only linearly with the number of stages. Its implementation, however, requires an exact solution of the proximal subproblem associated with the stage cost, and this subproblem often admits no closed-form solution. To resolve this issue, we develop an accelerated dynamic stochastic primal-dual (A-DSPD) framework that retains the recursive value-function structure and only requires the gradient of the stage cost. Within this framework, each stagewise saddle-point subproblem is solved by an inexact accelerated stochastic primal-dual method (iASPD), which applies Nesterov acceleration to the smooth stage cost and uses recursively generated stochastic $ε$-subgradients for the nonsmooth downstream value function. For the stochastic composite saddle-point problem, after $N$ primal-dual updates, iASPD yields optimality and feasibility bounds in which the contribution of the smooth stage cost decays at a faster rate $O(1/N^2)$, while the terms associated with stochastic $ε$-subgradients of the nonsmooth value function decay as $O(1/\sqrt{N})$, up to the prescribed oracle inexactness. We also develop bounds for the stochastic $ε$-subgradient estimators and the dual iterates, which yield complexity guarantees for A-DSPD. Specifically, A-DSPD computes an $ε$-solution using $O(1/ε^4)$ sampled scenarios and $O(1/ε^5)$ primal-dual updates for three-stage problems. We further extend the method and its convergence analysis to problems with an arbitrary finite number of stages. Numerical experiments on three-stage stochastic portfolio models illustrate the empirical advantage of the proposed method.

Explore related subjects

Keep this discovery

Explore connections, maps & timelines

BibTeXRIS

Bin Zhou, Zhiyun Guo, Hailin Sun, Qi Deng. 2026-11-10. An Accelerated Dynamic Stochastic Primal-Dual Method for Multi-stage Stochastic Optimization. https://searcharxiv.com/abs/2611.13905

Citations retain the original authors and DOI and include this site’s record URL. If this discovery service contributed to your research, you can also cite Searcharxiv by Indiciae: BibTeX · RIS.

KEEP EXPLORING

Related papers

First-order methods for stochastic and finite-sum convex optimization with deterministic constraints

In this paper, we study a class of stochastic and finite-sum convex optimization problems with deterministic constraints. Existing methods typically aim to find an $ε$-$expected\ feasible\ stochastic\ optimal$ solution, in which both the expected constraint violation and the expected optimality gap are within a prescribed tolerance $ε$. However, in many practical applications, constraints must be nearly satisfied with certainty, rendering such solutions potentially unsuitable due to the risk of substantial violations. To address this issue, we propose stochastic first-order methods for finding an $ε$-$surely\ feasible\ stochastic\ optimal$ ($ε$-SFSO) solution, where the constraint violation is deterministically bounded by $ε$ and the expected optimality gap is at most $ε$. Our methods apply an accelerated stochastic gradient (ASG) scheme or a modified variance-reduced ASG scheme $only\ once$ to a sequence of quadratic penalty subproblems with appropriately chosen penalty parameters. We establish first-order oracle complexity bounds for the proposed methods in computing an $ε$-SFSO solution. Preliminary numerical results are presented to illustrate the performance of the proposed methods.

math.OC↗

Attention's forward pass and Frank-Wolfe

We study the hardmax limit of self-attention dynamics for token embeddings obtained in the zero-temperature ($β\to+\infty$) regime, and relate it to the finite-$β$ setting. In this limit, the update rule can be viewed as a Frank-Wolfe step for a quadratic objective over the convex hull of the current token embeddings. When the key-query matrix is negative semidefinite, the method linearly contracts all tokens to a single cluster at the origin. When it is positive semidefinite, extending the hardmax rule to the entire convex hull induces a Voronoi diagram: vertices are stationary, interior points remain in their initial cells, and each token moves along a straight line toward its cell's vertex, yielding (super-)exponential convergence. As a byproduct, we also establish well-posedness of the associated ODE limit in this regime. Returning to the finite-$β$ regime, we model self-attention dynamics as a Markov chain and prove dynamic metastability: with high probability, interior tokens reach near-vertex configurations in a constant number of steps and remain within a small neighborhood for times that grow exponentially in the inverse temperature $β$, before ultimately collapsing to the origin. Thus, the hardmax dynamics accurately approximate the finite-$β$ process over exponentially long time horizons.

math.OC↗

Finite de Finetti for convex bodies and Polynomial Optimization

We study polynomial optimization over convex bodies with separately affine objectives and local affine equality and inequality constraints. We construct a convergent hierarchy of conic relaxations with explicit finite-level error bounds of order $n^{-1/2}$ at level $n$, complemented by a constructive rounding scheme that produces feasible points. The relaxation bounds and rounded objective values bracket the optimum at the same rate. The analytic ingredient is a finite de Finetti representation theorem for arbitrary proper cones: bipartite marginals of symmetric extensions in maximal tensor products are $O(n^{-1/2})$-close, in base norm, to points of the minimal tensor product. Its proof combines a uniform, monogamy-type bound on a variational mutual information based on Frenkel's relative entropy with classical conditioning and informationally complete measurements. We apply the framework to the assemblage relaxation of two-player non-local games in general probabilistic theories. For polytopic state spaces, the hierarchy consists of linear programs, and experiments on the Clauser--Horne--Shimony--Holt (CHSH) game provide a numerical proof of concept for the outer approximations and feasible rounding procedure.

math.OC↗