Portfolio Analysis Using Single Index, Multi-Index, and Constant Correlation Models: A Unified Treatment
提出一种适用于七种模型的简单通用算法，用于不允许卖空风险证券的最优投资组合选择，特别适用于无排序准则的多指数模型，并通过模拟研究证明其算法效率。
In this study a simple common algorithm which is applicable to seven models is proposed for optimal portfolio selection disallowing short sales of risky securities. The models considered in the analysis consist of a single index model, four multi-index models, and two constant correlation models. Unlike the previous approach, the proposed algorithm does not require explicit ranking of securities. Therefore, it is particularly useful for two multi-index models with orthogonal indices which do not provide any ranking criterion. Also, because of its algorithmic efficiency as demonstrated in a simulation study on models with multiple groups, the approach here can enhance their usefulness in portfolio analysis.