【摘要】 为提高遗传算法种群的多样性,在选择父辈串的时候通过综合考虑适应度值与彼此之间的海明距离确定选择机制,提出了基于适应度-距离(FD:Fitness-Distance)选择机制的遗传算法(GA:GeneticAlgorithm)。将该遗传算法与人工神经网络(ANN:ArtificialNeuralNetwork)技术相结合,应用于油气层的识别问题中。通过对实际样本的验证,获得了较好的效果,达到了快速识别油气层的目的,对两个实例的训练精度分别比标准遗传算法提高了22.6%和10.5%更多还原

【Abstract】 To improve the variation of the population in GA (Genetic Algorithm), the fitness values of the individuals and Haiming distances between each other are considered synthetically during the selection operation. The proposed method is named as Fitness-Distance Selection Mechanism Based GA and which is integrated with ANN (Artificial Neural Network) for the identification of oil gas layer. Through the verification to the real sample, good results are obtained, and the goal of fast identifying the oil gas layer is achieved. Of two cases the training accuracy of the proposed method are improved 22.6% and 10.5% than those of the standard GA.更多还原