摘要
该文研究带有工业约束和凹的交易费函数的离散单因素投资组合模型.与传统的投资组合模型不同的是,该模型中投资组合的决策变量是交易手数(整数),其最优化模型是一个非线性整数规划问题.为此提出了一个基于拉格朗日松弛和连续松弛的混合分枝定界算法,而且分别采用股票市场的真实数据和随机产生的数据来测试该算法的有效性.
We consider the discrete single-factor portfolio selection model with industry constraints and concave transaction cost, This model is a nonlinear integer programming problem. A hybrid branch-and-bound method based on Lagrangian relaxation and continuous relaxation is proposed for this model. Computational experiments are carried out with data both from the real-world stock market and generated randomly.
出处
《上海大学学报(自然科学版)》
CAS
CSCD
北大核心
2007年第6期736-740,共5页
Journal of Shanghai University:Natural Science Edition
基金
国家自然科学基金资助项目(1057111670518001)
关键词
金融优化
单因素模型
拉格朗日松弛
连续松弛
交易费
分枝定界法
portfolio optimization
single-factor model
Lagrangian relaxation
continuous relaxation
transaction cost
branch-and-bound method