针对中国证券市场现存的各种交易费用,建立一个更加符合实际的组合证券投资模型.该模型不仅继承了股票不可拆分、不能卖空等特点,而且完全反映了交易费用的核算情况;最后给出一个遗传算法结合动态罚函数求解的投资实例,计算结果证明了该模型及其求解方法的有效性和可操作性. Aiming at the various transaction costs existing in China’s securities market, this paper establishes a more realistic portfolio investment model, which not only inherits the characteristics of inseparable stocks and short sales, but also completely reflects the accounting of transaction costs. Finally, A genetic algorithm combined with a dynamic penalty function to solve the investment case, the results show that the model and its solution method is valid and operable.