年度	2023
计画类别	研究计画
计画名称	基于碎形布朗运动之多目标投资组合最佳化
参与人	陈耀宗
职称/担任之工作	主持人
计画期间	2023.08 ~ 2024.07
补助/委讬或合作机构	国家科学及技术委员会
摘要	Multi-objective optimization algorithms have been used in portfolio optimization for many years. However, when the optimized portfolio is tested in out-sample, the risk measured by the variance of return is not always optimal. That maybe because the assumption of a normal distribution of return on assets may be problematic. The return distribution simulated by the Fractal Brownian Motion (fBm) with long-term memory has a greater variance than the normal distribution, which is more suitable for capturing extreme return changes in the future. Therefore, if the return distribution simulated by the fBm is used as training data, the portfolio optimized by algorithms should better reflect the extreme changes in the financial market, which can make the theory of portfolio optimization and its function of guiding investment direction more reliable.
语言	中文