法人投資策略－以動態馬可夫模型分析

Abstract

本研究建立在法人進出可影響股價的不完全競爭市場上，並考慮具策略成本的策略，來分析法人如何利用策略以求取最大的利益。本研究使用動態馬可夫系統，以機率的觀念動態修正法人投資策略，建立法人投資理論的數學模型。在本文模型架構下，可得知在已知的機率結構下，法人的最佳投資策略及使用該策略後預期可獲得的報酬率。

Our research is based on a non-perfectly competitive market and supposed the trading strategies of the institution investors may affect underlying stock price with cost. While trying to find out the specific strategies to maximize the profit, we use the mathematical methods of Markov process and dynamic programming to develope an investment theory of institution investor. We take stock return levels as states and adopt the concepts of transition probability to revise the trading strategies. By using our model, institution investors may arrive at the optimal policy and the expected return from a given probability structure.