摘要： 为解决领域覆盖算法中覆盖中心的选取问题，引入遗传算法中的适应度函数，提出一种改进的覆盖算法。以覆盖样本数最多为目标设计适应度函数，通过计算每个样本的适应度值来搜索最优覆盖中心，采用神经网络和灵敏度相结合的方法计算输入因素对输出因素的决策权重。实验证明，该算法能保证覆盖的稳定性，且覆盖中心的个数较少。

关键词: 覆盖算法, 适应度函数, 灵敏度系数, 特征权重, 神经网络

Abstract: In order to solve the problem of selection of the center cover, this paper combines fitness function of the genetic algorithm with covering algorithm, and presents an improvement on neighborhood covering algorithm. Most samples are covered, and fitness function is designed. Decision weights are calculated by using neural networks and sensitivity of the method of combining. Experiments show this algorithm which searches the best coverage center in the sample space, ensures that the coverage is stable, and the number is less.

Key words: covering algorithm, fitness function, sensitivity coefficient, feature weight, neural network

中图分类号: 

• TP301.6