第三，研究了学习分类器系统在多机器人学习中的应用。
Thirdly, Learning Classifier system is applied to multi-robot system.
例如，某些基本的神经网络，它们的感知器只倾向于学习线形函数(通过划一条线可以把函数输入解析到分类系统中)。
For instance, a certain kind of basic neural network, the perceptron, is biased to learning only linear functions (functions with inputs that can be separated into classifications by drawing a line).
由确定性退火技术构造学习规则用于优化分类器参数，目的是减少分类误差以及待识别空间的系统熵。
Learning rules are constructed according to deterministic annealing to optimize classifier parameters, on purpose to reduce classification error and system entropy of the space to be identified.
基于竞争学习算法的模糊分类器确定系统的模糊空间和模糊规则数，并得出每个样本对每条规则的适用程度。
The fuzzy space structure of system and the number of fuzzy rules based on fuzzy competitive learning algorithm are determined and the fitness degree of each rule contrast to each sample is obtained.
对UCI机器学习数据库的实验证明，相对于其它方法，EPD方法对多分类器系统性能的预测能力更强。
The experiments on UCI Machine Learning Repository prove that, compared to existing measures, EPD shows stronger ability in predicting the performance of multiple classifier systems.
实验显示，相比单纯的基于机器学习的分类系统，这种组合型分类器产生了8%的性能提升。
The experiment result shows that there is 8% performance improvement compared with the single classifying method based on machine learning.
实验显示，相比单纯的基于机器学习的分类系统，这种组合型分类器产生了8%的性能提升。
The experiment result shows that there is 8% performance improvement compared with the single classifying method based on machine learning.
应用推荐