由於投資組合(Portfolio)理論中，Markowitz Model只考量了個別資產於投資組合中的權重，但無法決定現金保留比例的問題；而投資組合保險(Portfolio Insurance)只考量到了資金保留的問題，但無法決定風險性資產中各個標的物之權重。 本研究提出一個新的表達方式，將傳統動態投資組合保險中的兩種策略---固定比例投資組合保險策略(Constant Proportion Portfolio Insurance, CPPI)，與時間不變性投資組合保險策略(Time Invariant Portfolio Protection, TIPP)做模型上的擴充，使這兩個策略得以運用在投資組合之上。擴充後的策略模型，不僅同時考量了現金保留及資產權重的問題，且具有動態調整資產權重、及資金可相互流動的特性。 最後利用遺傳演算法之最佳化能力，搜尋新策略模型中最適合的乘數(Multiplier)及最大跌幅(Max DrawDown )，以求單位風險報酬率的最大化。 The Markiwotz Model only considers about the weight of each asset in a portfolio, but it doesn't consider preservation of cash. On the other hand, The Portfolio Insurance theory only considers preservation of cash. Our research extends two models of the traditional dynamic portfolio insurance --- Constant Proportion Portfolio Insurance (CPPI), and Time Invariant Portfolio Protection (TIPP), and apply those two new models to portfolio. Those two new models not only consider both the weight of each asset and preservation of cash, but also have the capability to dynamically adjust the weight of each asset. Finally, our research uses Genetic Algorithm to prove the performance of those two new models in Taiwan stock market, and the results of return and risks both better than the index.
