基于粒子群算法和BP神经网络的多因素林火等级预测模型
|
摘要
针对现有小尺度林火预测模型预测结果有效性、可扩展性等方面的不足,通过考虑多种火险因素,构建BP神经网络预测模型以提高预测精度,在此基础上借助粒子群算法加快BP神经网络收敛速度,进而提出一种混成的多因素森林火险等级预测模型particle swarm optimization based back-propagation neural network (PSO-BP)。所构建的预测模型,能够同时考虑气候因素(日最高气温、日平均气温、24 h降水量、连旱天数、日照时数、日平均相对湿度、日平均风速)、地形地貌因素(海拔、坡度、坡向、土壤含水量)、可燃物因素(植被类型、可燃物含水率、地被物载量)、人为因素(人口密度、距人类活动区域的距离) 16个变量。基于南京林业大学下蜀林场森林防火实验站传感器网络所采集的实际数据及现场测量数据,通过一组试验验证提出模型的有效性。结果表明:基于训练数据集及检验样本所构建的模型能够开展有效的火险等级预测;模型的计算复杂度较单独使用BP神经网络模型明显下降。
Forest fire destroys woodlands and forest resources,and emits massively greenhouse gases. It has been recognized as a series disaster for the sustainability of forests. Prediction of the risk rating for forest fire early warning has been a hot topic and widely investigated in recent years. To prevent forest fire and mitigate loss,an effective forest fire risk rating prediction desires for a real-time and region-related result. The relevant existing studies have proposed some forest fire risk rating prediction approaches. Nevertheless,the exiting approaches still cannot meet the previously mentioned forest fire prevention application requirements. Firstly,different models fit well for different regions( concerning climates and fuel types,etc.). Seldom approaches can scale for different regions. Secondly,the prediction results are split by days. To perform short-term multi-factor forest fire risk rating prediction and guarantee the accuracy and scalability of prediction model,we investigated the machine learning methodologies in this paper and proposed a particle swarm optimization( PSO)-based back-propagation( PSO-BP) neural network model. Specially,we incorporated the following 16 forest fire dangerous factors based on the existing reports in PSO-BP,i.e.,meteorological factors( e.g.,daily maximum air temperature,daily average air temperature,24-hour rainfall,the number of dry days,hours of sunshine,daily average relative humidity,and daily average wind velocity),terrain factors( e.g.,altitude,gradient,exposure,and soil water content),combustible factors( e.g.,types of vegetation,fuel moisture content,and content of forest litter),and human factor( e.g.,density of population,distance to the human activity sites). As for the PSOBP model,we firstly used BP neural network as the foundation technology to construct the prediction model. To guarantee the performance in terms of computational complexity of the prediction model,we then adopted PSO algorithm to speed up the convergence speed of BP neural network training process. We also presented a case study based on the data collected via the sensor-network-based forest fire prevention experimental station,Xiashu Forest,Nanjing Forestry University,China. The experimental results suggested that: 1) The PSO-BP prediction model constructed based on the training data set can effectively predict the forest fire danger ratings in the near-future for the study region; and 2) The efficiency in terms of computational complexity for training the prediction model using the PSO-BP is better than that of those models using solely BP neural network.
Forest fire destroys woodlands and forest resources,and emits massively greenhouse gases. It has been recognized as a series disaster for the sustainability of forests. Prediction of the risk rating for forest fire early warning has been a hot topic and widely investigated in recent years. To prevent forest fire and mitigate loss,an effective forest fire risk rating prediction desires for a real-time and region-related result. The relevant existing studies have proposed some forest fire risk rating prediction approaches. Nevertheless,the exiting approaches still cannot meet the previously mentioned forest fire prevention application requirements. Firstly,different models fit well for different regions( concerning climates and fuel types,etc.). Seldom approaches can scale for different regions. Secondly,the prediction results are split by days. To perform short-term multi-factor forest fire risk rating prediction and guarantee the accuracy and scalability of prediction model,we investigated the machine learning methodologies in this paper and proposed a particle swarm optimization( PSO)-based back-propagation( PSO-BP) neural network model. Specially,we incorporated the following 16 forest fire dangerous factors based on the existing reports in PSO-BP,i.e.,meteorological factors( e.g.,daily maximum air temperature,daily average air temperature,24-hour rainfall,the number of dry days,hours of sunshine,daily average relative humidity,and daily average wind velocity),terrain factors( e.g.,altitude,gradient,exposure,and soil water content),combustible factors( e.g.,types of vegetation,fuel moisture content,and content of forest litter),and human factor( e.g.,density of population,distance to the human activity sites). As for the PSOBP model,we firstly used BP neural network as the foundation technology to construct the prediction model. To guarantee the performance in terms of computational complexity of the prediction model,we then adopted PSO algorithm to speed up the convergence speed of BP neural network training process. We also presented a case study based on the data collected via the sensor-network-based forest fire prevention experimental station,Xiashu Forest,Nanjing Forestry University,China. The experimental results suggested that: 1) The PSO-BP prediction model constructed based on the training data set can effectively predict the forest fire danger ratings in the near-future for the study region; and 2) The efficiency in terms of computational complexity for training the prediction model using the PSO-BP is better than that of those models using solely BP neural network.
引文
[1]郭江.森林防火信息服务平台的设计与实现[D].北京:北京林业大学,2015.GUO J.Design and implementation of information service platform for forest fire prevention[D].Beijing:Beijing Forestry University,2015.
[2]TIAN X R,ZHAO F J,SHU L F,et al.Distribution characteristics and the influence factors of forest fires in China[J].Forest Ecology and Management,2013,310:460-467.DOI:10.1016/j.foreco.2013.08.025.
[3]POURTAGHI Z S,POURGHASEMI H R,ARETANO R,et al.Investigation of general indicators influencing on forest fire and its susceptibility modeling using different data mining techniques[J].Ecological Indicators,2016,64:72-84.DOI:10.1016/j.ecolind.2015.12.030.
[4]梁慧玲,郭福涛,苏漳文,等.基于随机森林算法的福建省林火发生主要气象因子分析[J].火灾科学,2015,24(4):191-200.DOI:10.3969/j.issn.1004-5309.2015.04.02.LIANG H L,GUO F T,SU Z W,et al.Analysis of meteorological factors on forest fire occurrence of Fujian based on random forest algorithm[J].Fire Safety Science,2015,24(4):191-200.
[5]苏漳文,刘爱琴,郭福涛,等.基于气象因子的福建省森林火险预测模型[J].森林与环境学报,2015,35(4):370-375.DOI:10.13324/j.cnki.jfcf.2015.04.013.SU Z W,LIU A Q,GUO F T,et al.Model to predict forest fire occurrence in Fujian Province based on meteorological factors[J].Journal of Forest and Environment,2015,35(4):370-375.
[6]王伟.基于森林火险指数的森林火险区划研究[J].林业资源管理,2015(4):79-83.DOI:10.13466/j.cnki.lyzygl.2015.04.014.WANG W.Study on forest fire risk zone based on forest fire risk index[J].Forest Resources Management,2015(4):79-83.
[7]曹姗姗.小尺度森林火险等级预测模型研究[D].北京:中国林业科学研究院,2014.CAO S S.Small-scale forest fire danger rating prediction[D].Beijing:Chinese Academy of Forestry,2014.
[8]SAKR G E,ELHAJJ I H,MITRI G.Efficient forest fire occurrence prediction for developing countries using two weather parameters[J].Engineering Applications of Artificial Intelligence,2011,24(5):888-894.DOI:10.1016/j.engappai.2011.02.017.
[9]张运林,张恒,金森.季节和降雨对细小可燃物含水率预测模型精度的影响[J].中南林业科技大学学报,2015,35(8):5-12.DOI:10.14067/j.cnki.1673-923x.2015.08.002.ZHANG Y L,ZHANG H,JIN S.Effects of season change and rainfall on forecast model accuracy of predicting fine fuels in forests in Pangu Forest Farm[J].Journal of Central South University of Forestry&Technology,2015,35(8):5-12.
[10]何泽能,唐晓萍,谭炳全.森林火险气象条件及等级预报初探---以重庆市沙坪坝区为例[J].灾害学,2013,28(2):46-50.DOI:10.3969/j.issn.1000-811X.2013.02.010HE Z N,TANG X P,TAN B Q.Study on meteorological condition and forecast of forest fire danger grading:a case study in Shapingba district of Chongqing[J].Journal of Catastrophology,2013,28(2):46-50.
[11]王磊,童振超,陈书林,等.森林防火智能化综合管理系统研究[J].林业经济,2014,36(12):113-116.DOI:10.13843/j.cnki.lyjj.2014.12.024.WANG L,TONG Z C,CHEN S L,et al.Study on the intelligent and integrated management information system for forest fire[J].Forestry Economics,2014,36(12):113-116.
[12]于淼.北京房山林火发生预测模型及小班火险等级区划研究[D].北京:北京林业大学,2016.YU M.The research of forest fire prediction model in Fangshan District,Beijing and sublot fire danger rating division[D].Beijing:Beijing Forestry University,2016.
[13]吴琳,张智光.我国“互联网+林业”的技术-产业-运作三维发展路径[J].世界林业研究,2018,31(4):1-7.DOI:10.13348/j.cnki.sjlyyj.2018.0063.y.WU L,ZHANG Z G.Research on three-dimensional technologyindustryoperation development path of“internet+forestry”in China[J].World Forestry Research,2018,31(4):1-7.
[14]张月琴,刘翔,孙先洋.一种改进的BP神经网络算法与应用[J].计算机技术与发展,2012,22(8):163-166.DOI:10.3969/j.issn.1673-629X.2012.08.042.ZHANG Y Q,LIU X,SUN X Y.An imporved algorithm of BPneural network and its application[J].Computer Technology and Development,2012,22(8):163-166.
[15]焦斌,叶明星.BP神经网络隐层单元数确定方法[J].上海电机学院学报,2013,16(3):113-116,124.DOI:10.3969/j.issn.2095-0020.2013.03.002.JIAO B,YE M X.Determination of hidden unit number in a BPneural network[J].Journal of Shanghai Dianji University,2013,16(3):113-116,124.
[16]郝娟.基于粒子群算法优化BP神经网络的SRM磁链模型[J].机械制造与自动化,2018,47(2):130-132.DOI:10.19344/j.cnki.issn1671-5276.2018.02.035.HAO J.SRM flux linkage model of optimizing BP neural network based on PSO algorithm[J].Machine Building&Automation,2018,47(2):130-132.
[17]YU F,XU X Z.A short-term load forecasting model of natural gas based on optimized genetic algorithm and improved BP neural network[J].Applied Energy,2014,134:102-113.DOI:10.1016/j.apenergy.2014.07.104.
[18]张德慧,张德育,刘清云,等.基于粒子群算法的BP神经网络优化技术[J].计算机工程与设计,2015,36(5):1321-1326.DOI:10.16208/j.issn1000-7024.2015.05.040.ZHANG D H,ZHANG D Y,LIU Q Y,et al.BP neural network optimized by improved PSO[J].Computer Engineering and Design,2015,36(5):1321-1326.
[19]LIU C J,DING W F,LI Z,et al.Prediction of high-speed grinding temperature of titanium matrix composites using BP neural network based on PSO algorithm[J].The International Journal of Advanced Manufacturing Technology,2017,89(5/6/7/8):2277-2285.DOI:10.1007/s00170-016-9267-z.
[20]江丽.基于粒子群与模拟退火算法的BP网络学习方法研究[D].合肥:安徽大学,2013.JIANG L.Research on BP neural network learning based on particle swarm optimization and simulated annealing algorithm[D].Hefei:Anhui University,2013.
[21]WANG L,ZHAO Q J,WEN Z M,et al.RAFFIA:short-term forest fire danger rating prediction via multiclass logistic regression[J].Sustainability,2018,10(12):4620.DOI:10.3390/su10124620.
[22]SAATY T L.Decision making with the analytic hierarchy process[J].International Journal of Services Sciences,2008,1(1):83.DOI:10.1504/ijssci.2008.017590.
[23]楼文高,乔龙.基于神经网络的金融风险预警模型及其实证研究[J].金融论坛,2011,16(11):52-61.DOI:10.16529/j.cnki.11-4613/f.2011.11.001.LOU W G,QIAO L.Early warning model of financial risks and empirical study based on neural network[J].Finance Forum,2011,16(11):52-61.
[2]TIAN X R,ZHAO F J,SHU L F,et al.Distribution characteristics and the influence factors of forest fires in China[J].Forest Ecology and Management,2013,310:460-467.DOI:10.1016/j.foreco.2013.08.025.
[3]POURTAGHI Z S,POURGHASEMI H R,ARETANO R,et al.Investigation of general indicators influencing on forest fire and its susceptibility modeling using different data mining techniques[J].Ecological Indicators,2016,64:72-84.DOI:10.1016/j.ecolind.2015.12.030.
[4]梁慧玲,郭福涛,苏漳文,等.基于随机森林算法的福建省林火发生主要气象因子分析[J].火灾科学,2015,24(4):191-200.DOI:10.3969/j.issn.1004-5309.2015.04.02.LIANG H L,GUO F T,SU Z W,et al.Analysis of meteorological factors on forest fire occurrence of Fujian based on random forest algorithm[J].Fire Safety Science,2015,24(4):191-200.
[5]苏漳文,刘爱琴,郭福涛,等.基于气象因子的福建省森林火险预测模型[J].森林与环境学报,2015,35(4):370-375.DOI:10.13324/j.cnki.jfcf.2015.04.013.SU Z W,LIU A Q,GUO F T,et al.Model to predict forest fire occurrence in Fujian Province based on meteorological factors[J].Journal of Forest and Environment,2015,35(4):370-375.
[6]王伟.基于森林火险指数的森林火险区划研究[J].林业资源管理,2015(4):79-83.DOI:10.13466/j.cnki.lyzygl.2015.04.014.WANG W.Study on forest fire risk zone based on forest fire risk index[J].Forest Resources Management,2015(4):79-83.
[7]曹姗姗.小尺度森林火险等级预测模型研究[D].北京:中国林业科学研究院,2014.CAO S S.Small-scale forest fire danger rating prediction[D].Beijing:Chinese Academy of Forestry,2014.
[8]SAKR G E,ELHAJJ I H,MITRI G.Efficient forest fire occurrence prediction for developing countries using two weather parameters[J].Engineering Applications of Artificial Intelligence,2011,24(5):888-894.DOI:10.1016/j.engappai.2011.02.017.
[9]张运林,张恒,金森.季节和降雨对细小可燃物含水率预测模型精度的影响[J].中南林业科技大学学报,2015,35(8):5-12.DOI:10.14067/j.cnki.1673-923x.2015.08.002.ZHANG Y L,ZHANG H,JIN S.Effects of season change and rainfall on forecast model accuracy of predicting fine fuels in forests in Pangu Forest Farm[J].Journal of Central South University of Forestry&Technology,2015,35(8):5-12.
[10]何泽能,唐晓萍,谭炳全.森林火险气象条件及等级预报初探---以重庆市沙坪坝区为例[J].灾害学,2013,28(2):46-50.DOI:10.3969/j.issn.1000-811X.2013.02.010HE Z N,TANG X P,TAN B Q.Study on meteorological condition and forecast of forest fire danger grading:a case study in Shapingba district of Chongqing[J].Journal of Catastrophology,2013,28(2):46-50.
[11]王磊,童振超,陈书林,等.森林防火智能化综合管理系统研究[J].林业经济,2014,36(12):113-116.DOI:10.13843/j.cnki.lyjj.2014.12.024.WANG L,TONG Z C,CHEN S L,et al.Study on the intelligent and integrated management information system for forest fire[J].Forestry Economics,2014,36(12):113-116.
[12]于淼.北京房山林火发生预测模型及小班火险等级区划研究[D].北京:北京林业大学,2016.YU M.The research of forest fire prediction model in Fangshan District,Beijing and sublot fire danger rating division[D].Beijing:Beijing Forestry University,2016.
[13]吴琳,张智光.我国“互联网+林业”的技术-产业-运作三维发展路径[J].世界林业研究,2018,31(4):1-7.DOI:10.13348/j.cnki.sjlyyj.2018.0063.y.WU L,ZHANG Z G.Research on three-dimensional technologyindustryoperation development path of“internet+forestry”in China[J].World Forestry Research,2018,31(4):1-7.
[14]张月琴,刘翔,孙先洋.一种改进的BP神经网络算法与应用[J].计算机技术与发展,2012,22(8):163-166.DOI:10.3969/j.issn.1673-629X.2012.08.042.ZHANG Y Q,LIU X,SUN X Y.An imporved algorithm of BPneural network and its application[J].Computer Technology and Development,2012,22(8):163-166.
[15]焦斌,叶明星.BP神经网络隐层单元数确定方法[J].上海电机学院学报,2013,16(3):113-116,124.DOI:10.3969/j.issn.2095-0020.2013.03.002.JIAO B,YE M X.Determination of hidden unit number in a BPneural network[J].Journal of Shanghai Dianji University,2013,16(3):113-116,124.
[16]郝娟.基于粒子群算法优化BP神经网络的SRM磁链模型[J].机械制造与自动化,2018,47(2):130-132.DOI:10.19344/j.cnki.issn1671-5276.2018.02.035.HAO J.SRM flux linkage model of optimizing BP neural network based on PSO algorithm[J].Machine Building&Automation,2018,47(2):130-132.
[17]YU F,XU X Z.A short-term load forecasting model of natural gas based on optimized genetic algorithm and improved BP neural network[J].Applied Energy,2014,134:102-113.DOI:10.1016/j.apenergy.2014.07.104.
[18]张德慧,张德育,刘清云,等.基于粒子群算法的BP神经网络优化技术[J].计算机工程与设计,2015,36(5):1321-1326.DOI:10.16208/j.issn1000-7024.2015.05.040.ZHANG D H,ZHANG D Y,LIU Q Y,et al.BP neural network optimized by improved PSO[J].Computer Engineering and Design,2015,36(5):1321-1326.
[19]LIU C J,DING W F,LI Z,et al.Prediction of high-speed grinding temperature of titanium matrix composites using BP neural network based on PSO algorithm[J].The International Journal of Advanced Manufacturing Technology,2017,89(5/6/7/8):2277-2285.DOI:10.1007/s00170-016-9267-z.
[20]江丽.基于粒子群与模拟退火算法的BP网络学习方法研究[D].合肥:安徽大学,2013.JIANG L.Research on BP neural network learning based on particle swarm optimization and simulated annealing algorithm[D].Hefei:Anhui University,2013.
[21]WANG L,ZHAO Q J,WEN Z M,et al.RAFFIA:short-term forest fire danger rating prediction via multiclass logistic regression[J].Sustainability,2018,10(12):4620.DOI:10.3390/su10124620.
[22]SAATY T L.Decision making with the analytic hierarchy process[J].International Journal of Services Sciences,2008,1(1):83.DOI:10.1504/ijssci.2008.017590.
[23]楼文高,乔龙.基于神经网络的金融风险预警模型及其实证研究[J].金融论坛,2011,16(11):52-61.DOI:10.16529/j.cnki.11-4613/f.2011.11.001.LOU W G,QIAO L.Early warning model of financial risks and empirical study based on neural network[J].Finance Forum,2011,16(11):52-61.