arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 1804.00636math.OCstat.APstat.MEstat.ML
收藏

凸风险测度的递归Optimization:均值-半恒等式模型

Recursive Optimization of Convex Risk Measures: Mean-Semideviation Models

Dionysios S. Kalogerias, Warren B. Powell

摘要:

我们开发递归的,数据驱动的,随机次梯度方法优化一个新的,通用的,和应用驱动类凸风险的措施,这里被称为平均半理想,严格推广众所周知的和流行的平均上半理想。我们引入MESSAGEp算法,这是一个有效的组合次梯度迭代求解凸均值半理想风险厌恶问题的最优性。我们分析了MESSAGEp算法的渐近行为下的灵活和结构开发的问题假设。特别是:1)在适当的步长规则下,我们建立了MESSAGEp算法在强技术意义下的路径收敛性,证明了其渐近一致性。2)假设一个强凸成本,我们表明,对于固定的半行列式阶$p>1$和$ε\in\left[0,1\right)$,MESSAGEp算法实现了平方${\cal L}_{2}$解决方案次优率的顺序的${\cal O}(n^{-\left(1-ε\right)/2})$迭代,其中,$ε>0$,路径收敛同时保证。这个结果建立了一个任意接近${\cal O}(n^{-1/2})$的阶数,同时确保了强稳定的路径操作。对于$p\equiv1$,速率顺序提高到${\cal O}(n^{-2/3})$,这也足以进行路径收敛,并匹配以前的结果。3)同样,在一般情况下的凸成本,我们表明,对于任何$ε\in\left[0,1\right)$,MESSAGEp算法与平滑达到${\cal L}_{1}$目标次优率的顺序的${\cal O}(n^{-\left(1-ε\right)/\left(4\bf{1}_{\left\{ p>1\right\} }+4\right)})$迭代。如果是$p\equiv1$,则该结果提供了${\cal O}(n^{-1/4})$的最大速率,如果是$p>1$,则该结果提供了${\cal O}(n^{-1/8})$的最大速率,也与现有技术匹配。

英文摘要:

We develop recursive, data-driven, stochastic subgradient methods for optimizing a new, versatile, and application-driven class of convex risk measures, termed here as mean-semideviations, strictly generalizing the well-known and popular mean-upper-semideviation. We introduce the MESSAGEp algorithm, which is an efficient compositional subgradient procedure for iteratively solving convex mean-semideviation risk-averse problems to optimality. We analyze the asymptotic behavior of the MESSAGEp algorithm under a flexible and structure-exploiting set of problem assumptions. In particular: 1) Under appropriate stepsize rules, we establish pathwise convergence of the MESSAGEp algorithm in a strong technical sense, confirming its asymptotic consistency. 2) Assuming a strongly convex cost, we show that, for fixed semideviation order $p>1$ and for $ε\in\left[0,1\right)$, the MESSAGEp algorithm achieves a squared-${\cal L}_{2}$ solution suboptimality rate of the order of ${\cal O}(n^{-\left(1-ε\right)/2})$ iterations, where, for $ε>0$, pathwise convergence is simultaneously guaranteed. This result establishes a rate of order arbitrarily close to ${\cal O}(n^{-1/2})$, while ensuring strongly stable pathwise operation. For $p\equiv1$, the rate order improves to ${\cal O}(n^{-2/3})$, which also suffices for pathwise convergence, and matches previous results. 3) Likewise, in the general case of a convex cost, we show that, for any $ε\in\left[0,1\right)$, the MESSAGEp algorithm with iterate smoothing achieves an ${\cal L}_{1}$ objective suboptimality rate of the order of ${\cal O}(n^{-\left(1-ε\right)/\left(4\bf{1}_{\left\{ p>1\right\} }+4\right)})$ iterations. This result provides maximal rates of ${\cal O}(n^{-1/4})$, if $p\equiv1$, and ${\cal O}(n^{-1/8})$, if $p>1$, matching the state of the art, as well.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑