基于量子行为花朵授粉算法优化LSTM模型  

Optimizing LSTM Model Based on Quantum-Inspired Flower Pollination Algorithm

作　　者：李汝嘉 贺壹婷 季荣彪 李亚东 孙晓海 陈娇娇 吴叶辉 王灿宇[1] LI Rujia;HE Yiting;JI Rongbiao;LI Yadong;SUN Xiaohai;CHEN Jiaojiao;WU Yehui;WANG Canyu(College of Big Data,Yunnan Agricultural University,Kunming 650201,China)

机构地区：[1]云南农业大学大数据学院,昆明650201

出　　处：《吉林大学学报（理学版）》2024年第5期1163-1178,共16页Journal of Jilin University:Science Edition

基　　金：云南省重大科技专项基金(批准号:202302AE09002003)。

摘　　要：针对传统花朵授粉算法(flower pollination algorithm, FPA)受初始参数影响较大、且易陷入局部最优解或算法无法收敛等问题,提出一种基于量子行为的花朵授粉算法(quantum-inspired flower pollination algorithm, QFPA).通过引入量子系统到FPA中,使授粉过程中的搜索更高效,从而提高全局搜索能力.此外,还引入轨迹分析,使种群能更好地逃离局部最优解,进一步降低误差.为验证该方法的有效性,先通过选定的几个基准函数对QFPA进行评估,然后采用评估效果最好的QFPA对长短期记忆网络(LSTM)模型超参数进行寻优,最后在用自适应噪声完备集合经验模态分解(complete ensemble empirical mode decomposition with adaptive noise, CEEMDAN)算法去除噪声后的空气质量数据集上进行实验,并与其他几种常用的优化算法进行对比.实验结果表明:QFPA提高了优化算法的全局搜索能力和收敛性;QFPA-LSTM模型增强了长时间序列数据预测的准确性和效率,该模型预测的均方根误差为10.93μg/m^(3),为实际应用中的空气质量预测提供了可靠的解决方案.Aiming at the problem that the traditional flower pollination algorithm(FPA)was significantly affected by initial parameters and prone to local optima or convergence failures,we proposed a quantum-inspired flower pollination algorithm(QFPA).By incorporating quantum systems into the FPA,the pollination search process was made more efficient,thereby improving global search capabilities.Additionally,trajectory analysis was employed to better enable the population to escape from local optima and further reduce errors.In order to verify the effectiveness of the method,firstly,the QFPA was evaluated using selected benchmark functions.Secondly,the best evaluated QFPA was used to optimize the hyperparameters of the long short-term memory network(LSTM)model.Finally,the experiments were conducted on an air quality dataset after removing noise using the complete ensemble empirical mode decomposition with adaptive noise(CEEMDAN)algorithm,and compared with several other commonly used optimization algorithms.The experimental results show that QFPA enhances the global search capability and convergence properties of optimization algorithms.The QFPA-LSTM model improves the accuracy and efficiency of long-term time series predictions,with a root mean square error of 10.93μg/m^(3),thus providing a reliable solution for air quality prediction in practical applications.

关 键 词：花朵授粉算法 量子行为花朵授粉算法 CEEMDAN算法 LSTM模型 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]