【摘要】 考虑到中国证券交易的限制规定及现实投资者并非完全理性的决策行为,给出了组合投资收益-损失风险双目标概率准则的整数规划模型。通过证券收益经验分布,应用分层抽样的随机模拟,并结合动态变化算子的遗传算法,构造GASS II遗传模拟混合算法,进行概率准则模型的优化求解。股票相关性由其秩相关系数给出,算法将秩和区间划分相联系,指导分层抽样。GASS II算法能有效刻画收益分布的“高峰厚尾”,激发遗传算法的隐含并行搜索特性,避免早熟现象,提高寻优效率与精度。最后给出了一个投资组合实证分析算例的收益-损失风险有效前沿。更多还原

【Abstract】 Considering portfolio selection issue with the realistic environments of the stock exchange market in China,this paper presents probability criterion model of portfolio selection.The experimental distribution acquired on real history data is adopted to approach the real distribution and stratified sampling technique is employed.GASS II(genetic algorithm with stochastic simulation of stratified sampling intelligently integrated) is constructed to solve the probability criterion model.A demonstration with bearable return-loss risk efficient frontier is provided.更多还原