智能办公环境中多Agent模糊Q学习研究

目前有关智能办公环境的研究忽视了对建筑环境的考虑，为了给工作人员提供一个节能、舒适、便捷的办公环境，研究了智能办公环境无线网络系统的学习方法，即为系统建立了多智能体（Agent）模型，基于该模型提出了改进的模糊Q学习算法，用于学习用户的行为，以自动控制环境中执行器的动作。通过对环境温度学习的实验，比较了该算法和普通Q学习的MSE（Mean Square Error）值，实验结果表明提出的算法较优。

Research on multi-Agent fuzzy Q-learning in intelligent office environment

Presently,the research on intelligent office environment neglects the consideration of the building environment.In order to provide an energy-efficient,comfortable,convenient office environment for staffs,the learning method of wireless network system for intelligent office environment is studied.Namely,a multi-Agent model is established for the system,and on basis of which the improved fuzzy Q-learning algorithm is advanced,which is used to study user behavior in the environment,in order to control user action automatically.By the experiment on the ambient temperature study,the MSE(Mean Square Error)values of the improved fuzzy Q-learning algorithm and the ordinary Q are compared,and the result shows that the proposed algorithm is better.