基于贝叶斯优化梯度提升回归树的智慧酒店室内感应灯光设计  

Design of Indoor Induction Lighting for Smart Hotels Based on Bayesian Optimization Gradient Boosting Regression Tree

作　　者：王静[1] 孟梅林[2] WANG Jing;MENG MeiLin(College of Landscape and Horticulture,Wuhu Institute of Technology,Wuhu Anhui 241003,China;School of Design,Anhui Polytechnic University,Wuhu Anhui 241000,China)

机构地区：[1]芜湖职业技术学院园林园艺学院,安徽芜湖241003 [2]安徽工程大学设计学院,安徽芜湖241000

出　　处：《佳木斯大学学报(自然科学版)》2024年第12期158-161,165,共5页Journal of Jiamusi University:Natural Science Edition

基　　金：2020年安徽省教育厅教学示范课(2020SJJXSFK2533);2021年芜湖职业技术学院人文重点科研项目(wzyrwzd202114);2021年芜湖职业技术学院自然科学科研项目(wzyrwzd202111);2022年芜湖职业技术学院示范基层教研室项目(2022sfjys05)。

摘　　要：针对自然光非线性变化导致酒店环境不能随时保持较好照度的问题,研究提出了一种基于贝叶斯优化梯度提升回归树的室内日光照度预测方法。该方法采用高斯函数对单个照明装置产生的照度峰值进行预测,然后引入了高斯混合误差模型来补偿预测误差,最后通过贝叶斯优化梯度提升回归树预测室内日光照度。实验结果表明,在不同人员数量下,贝叶斯优化梯度提升回归树预测模型的预测误差值为32 Lux,47 Lux,31 Lux和17 Lux。所提出的预测模型能够对室内日光照度进行准确预测,从而实现感应灯光的智能动态调控,保证室内光照明系统的高效运行,为照明系统感应灯光调控提供决策参考。A method for predicting indoor sunlight illuminance based on Bayesian optimization gradient boosting regression tree is proposed to address the problem of non-linear changes in natural light that prevent hotel environments from maintaining good illuminance at all times.This method uses a Gaussian function to predict the peak illuminance generated by a single lighting device,and then introduces a Gaussian mixture error model to compensate for prediction errors.Finally,the Bayesian optimization gradient boosting regression tree is used to predict indoor sunlight illuminance.The experimental results show that under different numbers of personnel,the prediction error values of the Bayesian optimized gradient boosting regression tree prediction model are 32 Lux,47 Lux,31 Lux,and 17 Lux.The proposed prediction model can accurately predict indoor sunlight illumination,thereby achieving intelligent dynamic control of induced lighting,ensuring the efficient operation of indoor lighting systems,and providing decision-making reference for induced lighting control in lighting systems.

关 键 词：智慧酒店 贝叶斯优化 梯度提升回归树 感应灯光 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]