尝试用遗传算法与神经网络结合，对港口投资的可行性进行衡量，并建立了相应的投资决策模型，结合实例计算分析得到满意的结果。该方法有效地解决了传统ＢＰ算法对多峰值问题存在收敛于局部非劣解的弊病，而且大大加快了收敛的速度。 Attempt to combine the genetic algorithm with neural network to measure the feasibility of the port investment, and set up the corresponding investment decision-making model, with examples of calculation and analysis to obtain satisfactory results. This method effectively solves the shortcomings that the traditional BP algorithm converges to local non-inferior solutions to multi-peak problems, and greatly speeds up the convergence.