多阶段投资决策问题的一种智能化求解方法

对多阶段投资决策问题进行研究，建立一种极小化跟踪投资回报率与目标回报率偏差的多阶段投资决策模型，并将随机模拟、遗体算法和神经网络集成在动态规划之中，设计给出一种智能化的求解方法，能求得反馈形式的最优投资策略。本文给出的方法克服了传统求解方法的局限性，具有现实意义，经算例仿真验证了算法的可行性。

An Intelligent Solving Method for Multi-Period Investment Decision

In this paper, the problem of multi period investment decision is investigated. The mathematical model which to minimize the tracking error of return rate and expect rate is established. The Monte Carlo Simulation, Genetic Algorithms and artificial neural networks are integrated in the dynamic programming. In order to overcome the limitation of the traditional method of solution, an intelligent solving method for the problem of multi period investment decision is devised. The optimal feedback strategies can be obtained by using this solving method. Some examples show that the intelligent solving method is feasible.