基于混合元启发式算法优化神经网络的热舒适度预测模型

Thermal Comfort Prediction Model Based on Improved Hybrid Meta-heuristic Algorithm for Neural Network

热舒适度指标的计算过程具有多参数、非线性、高复杂度等特点,致使空调的实时控制器无法直接使用。针对这一问题,提出了一种基于混合元启发式算法神经网络的热舒适度预测模型:GACS神经网络。其中,GACS算法是一类融合了遗传算法和布谷鸟搜索的混合元启发式算法。仿真实验表明:与遗传算法相比,GACS算法在全局搜索能力方面得到了极大提升,其优化出的GACS神经网络具有更高的预测精度。

The calculation process of thermal comfort index has the characteristics of multi-parameters,non-linearity and high complexity,which makes the real-time controller of air conditioning unable to be used directly.To solve this problem,a thermal comfort prediction model based on hybrid meta-heuristic algorithm and neural network is proposed,i.e.,GACS neural network.GACS is a hybrid meta-heuristic algorithm,it combines genetic algorithm and Cuckoo search.The simulation results show that compared with genetic algorithm,GACS algorithm has greatly improved the global search ability,and the optimized GACS neural network has higher prediction accuracy.