投资组合管理

• 交叉列表
• 替换

查看 最近的 文章

显示 2025年08月07日， 星期四 新的列表

交叉提交 (展示 1 之 1 条目 )

[1] arXiv:2508.03704 (交叉列表自 q-fin.PM) [中文pdf, pdf, html, 其他]

标题： 基于等相关性投资组合策略的新型风险度量用于投资组合优化 显示英文标题

标题： Novel Risk Measures for Portfolio Optimization Using Equal-Correlation Portfolio Strategy

Biswarup Chakraborty

主题： 投资组合管理 (q-fin.PM) ; 应用 (stat.AP)

组合优化长期以来一直由基于协方差的策略主导，例如马科维茨均值-方差框架。 然而，这些方法往往无法确保资产之间的风险结构平衡，导致集中在少数证券上。 在本文中，我们引入了基于等相关性组合策略的新风险度量，旨在构建每个资产与整体组合收益保持相等相关性的组合。 我们制定了一个数学优化框架，明确控制组合范围内的相关性，同时保持理想的风险与回报权衡。 所提出的模型通过历史股票市场数据进行了实证验证。 我们的研究结果表明，通过这种方法构建的组合在各种市场条件下表现出更好的风险分散和更稳定的回报。 该方法为传统的多样化技术提供了一个有吸引力的替代方案，并对机构投资者、资产管理者和量化交易策略具有实际意义。 显示英文摘要

Portfolio optimization has long been dominated by covariance-based strategies, such as the Markowitz Mean-Variance framework. However, these approaches often fail to ensure a balanced risk structure across assets, leading to concentration in a few securities. In this paper, we introduce novel risk measures grounded in the equal-correlation portfolio strategy, aiming to construct portfolios where each asset maintains an equal correlation with the overall portfolio return. We formulate a mathematical optimization framework that explicitly controls portfolio-wide correlation while preserving desirable risk-return trade-offs. The proposed models are empirically validated using historical stock market data. Our findings show that portfolios constructed via this approach demonstrate superior risk diversification and more stable returns under diverse market conditions. This methodology offers a compelling alternative to conventional diversification techniques and holds practical relevance for institutional investors, asset managers, and quantitative trading strategies.

替换提交 (展示 1 之 1 条目 )

[2] arXiv:2307.07694 (替换) [中文pdf, pdf, html, 其他]

标题： 深度强化学习算法在投资组合优化中的评估 显示英文标题

标题： Evaluation of Deep Reinforcement Learning Algorithms for Portfolio Optimisation

Chung I Lu

主题： 计算工程、金融与科学 (cs.CE) ; 投资组合管理 (q-fin.PM)

我们在使用模拟数据的组合优化任务上评估了基准深度强化学习算法。 生成数据的模拟器基于具有Bertsimas-Lo市场影响模型的相关几何布朗运动。 使用Kelly准则（对数效用）作为目标，我们可以解析地推导出没有市场影响的最优策略，作为在包含市场影响时衡量性能的上限。 我们发现离策略算法DDPG、TD3和SAC由于奖励的噪声而无法学习正确的$Q$-函数，因此表现不佳。 使用广义优势估计的在策略算法PPO和A2C能够处理噪声并推导出接近最优的策略。 发现PPO的裁剪变体在防止策略在收敛后偏离最优策略方面很重要。 在一个更具挑战性的环境中，我们有GBM参数的制度变化，我们发现PPO结合隐藏马尔可夫模型来学习和预测制度上下文，能够学习到适应每个制度的不同策略。 总体而言，我们发现这些算法的样本复杂度对于使用真实数据的应用来说太高，即使在最简单的设置中也需要超过2m步来学习一个良好的策略，这相当于几乎8000年的每日价格。 显示英文摘要

We evaluate benchmark deep reinforcement learning algorithms on the task of portfolio optimisation using simulated data. The simulator to generate the data is based on correlated geometric Brownian motion with the Bertsimas-Lo market impact model. Using the Kelly criterion (log utility) as the objective, we can analytically derive the optimal policy without market impact as an upper bound to measure performance when including market impact. We find that the off-policy algorithms DDPG, TD3 and SAC are unable to learn the right $Q$-function due to the noisy rewards and therefore perform poorly. The on-policy algorithms PPO and A2C, with the use of generalised advantage estimation, are able to deal with the noise and derive a close to optimal policy. The clipping variant of PPO was found to be important in preventing the policy from deviating from the optimal once converged. In a more challenging environment where we have regime changes in the GBM parameters, we find that PPO, combined with a hidden Markov model to learn and predict the regime context, is able to learn different policies adapted to each regime. Overall, we find that the sample complexity of these algorithms is too high for applications using real data, requiring more than 2m steps to learn a good policy in the simplest setting, which is equivalent to almost 8,000 years of daily prices.