项目名称: 资产数目与投资周期带有基数约束的投资组合优化
项目编号: No.71201102
项目类型: 青年科学基金项目
立项/批准年度: 2013
项目学科: 管理科学与工程
项目作者: 高建军
作者单位: 上海交通大学
项目金额: 20万元
中文摘要: 以缩小传统投资组合模型和风险管理模型与实践应用的差距为目标,本项目研究两类有重要应用的带有基数约束的金融优化问题:1)由于交易费用的影响,投资组合模型需要加入对资产数目的限制,这使得原本容易求解模型变为NP-Hard的优化问题。本项目在均值-方差和均值-CVaR框架下研究资产数目带有基数约束的投资组合问题。利用问题结构上的特点,建立有效的松弛方法并结合优化领域的最新锥优化技术和割平面技术构造高效的数值求解算法,并使用市场的真实数据进行实证研究和分析。2)在一个长期多阶段的投资过程中,由于管理风险资产的成本影响,投资者往往不会长期持有风险资产。因此在动态投资组合管理的模型中需要考虑投资周期的基数限制。在多阶段均值-方差投资组合优化模型的意义下,研究如何利用动态规划以及其逼近方法寻找最优和近似最优的投资策略,并在此基础上分析基数约束和投资成本对投资策略的影响以及与市场时机选择问题的关系。
中文关键词: 投资组合优化;基数约束;半正定规划;随机动态规划;
英文摘要: To reduce the gap between the traditional portfolio management model and the real financial practice, this project studies two important financial optimization problems: 1) Due to the transaction cost, the number of assets should be limited in the portfolio selection model, which makes the optimization problem to be NP-Hard in general. This research studies the cardinality constrained portfolio selection problem under the framework of mean-variance and mean-CVaR. By exploring the special structure of these problems, we develop tight relaxation schemes and develop more efficient solution algorithms using conic optimization techniques and cutting plane methods. Empirical study will be also carried out to evaluate the proposed algorithms using real market data. 2) In a long-time multiperiod investment, the investor will not hold the risky assets in all time periods, due to the management cost of the risky asset. Thus, the cardinality constraint on the total number of periods holding the risky assets should be included in the dynamic portfolio management model. Under the mean-variance and utility maximization model, we explore how to use the dynamic programming to achieve the optimal and sub-optimal investment strategy. Based on such investment strategy, we analyze the impact of cardinality constraint and the conne
英文关键词: Portfolio optimization;Cardinality constraint;Semidefinite Programming;Stochastic dynamic programming;