ACO-BP模型在某实验室热舒适预测中的研究

Research ACO-BP in a Laboratory Model to Predict the Thermal Comfort

以热舒适指标PMV作为空调控制系统的控制目标,能够很大程度上实现舒适与节能的统一。基于此研究了一种蚁群神经网络预测分类模型,并给出了详细的设计步骤和部分Matlab设计代码,最后采用某大学实验室数据库中的夏天数据集进行了验证。结果表明,采用蚁群算法对BP神经网络进行整定后,不仅克服了BP算法容易陷入局部最优的缺点,也加快了蚂蚁的收敛速度,提高了热舒适预测分类的准确性。

Taking thermal comfort index PMV as a control target of air-conditioning control system, comfort and energy efficiency can be achieved largely unified. An ant colony neural network predictive classification model was developed,and a detailed design steps and some Matlab design code was given. Finally, the model was verified by a university laboratory database summer datasets. The results showed that the use of ant colony algorithm BP neural network tuning,BP algorithm not only overcome the shortcomings easy to fall into local optimum, but also accelerate the convergence speed of ants to improve the accuracy of prediction classification.