Integration of time series forecasting in a dynamic decision support system for multiple reservoir management to conserve water sources

In most of arid and semi-arid regions, there are limited sources of available fresh water for different domestic and environmental demands. Strategic and parsimonious fresh water-use in water-scarce areas such as Southern New Mexico is crucially important. Elephant Butte and Caballo reservoirs are two integrated reservoirs in this region that provide water supply for many water users in downstream areas. Since Elephant Butte Reservoir is in a semi-arid region, it would be rational to utilize other energy sources such as wind energy to produce electricity and use the water supply to other critical demands in terms of time and availability. This study develops a strategy of optimal management of two integrated reservoirs to quantify the savable volume of water sources through optimal operation management. To optimize operations for the Elephant Butte and Caballo reservoirs as an integrated reservoir operation in New Mexico, the authors in this case study utilized two autoregressive integrated moving average models, one non-seasonal (daily, ARIMA model) and one seasonal (monthly, SARIMA model), to predict daily and monthly inflows to the Elephant Butte Reservoir. The coefficient of determination between predicted and observed daily values and the normalized mean of absolute error (NMAE) were 0.97 and 0.09, respectively, indicating that the daily ARIMA prediction model was significantly reliable and accurate for a univariate based streamflow forecast model. The developed time series prediction models were incorporated in a decision support system, which utilizes the predicted values for a day and a month ahead and leads to save significant amount of water volume by providing the optimal release schedule from Elephant Butte into the Caballo Reservoir. The predicted daily and monthly values from the developed ARIMA prediction models were integrated successfully with the dynamic operation model, which provides the optimal operation plans. The optimal operation plan significantly minimizes the total evaporation loss from both reservoirs by providing the optimal storage levels in both reservoirs. The saved volume of the water would be considered as a significant water supply for environmental conservation actions in downstream of the Caballo Reservoir. Providing an integrated optimal management plan for two reservoirs led to save significant water sources in a region that water shortage has led to significant environmental consequences. Finally, since the models are univariate, they demonstrate an approach for reliable inflow prediction when information is limited to only streamflow values. We find that hydroelectric power generation forces the region to lose significant amount of water to evaporation and therefore hinder the optimal use of freshwater. Based on these findings, we conclude that a water scarce region like Southern New Mexico should gain independence from hydroelectric power and save the freshwater for supporting ecosystem services and environmental purposes.     

Time series study of the impact of serious fires on fire occurrence statistics in cities of Jiangsu

In this paper, the time series method is used to analyze the effect of fire disasters and serious fires on monthly fire occurrence statistics in 10 cities from January 1997 to December 2001 in Jiangsu province. After removal of outliers in the irregular component of the original data series and seasonal adjustments, the intervention model is applied to quantify the effect of the fire disasters and serious fires on the total monthly fire statistics in the cities of Jiangsu province. The results show that the impact on monthly fires lasts for 3 months for eight cities and 2 months for two cities. Other phenomena are also detected in these 10 cities. Finally, some explanations for these phenomena are proposed.     

基于小波分析的LS-SVM—ARIMA组合模型的黄金价格预测

黄金作为一种特殊的金融商品,其价格受国际原油、美元汇率、通货膨胀等多种因素的影响,波动性强。使用单一模型进行黄金价格预测通常效果不佳,只有充分考虑价格变化的各个方面才能更加准确地预测黄金价格。应用小波分析将黄金价格分解为4个不同变化趋势,应用LS-SVM与ARIMA模型对不同变化趋势进行建模预测,并重构黄金价格组合预测的结果。实证结果表明,该组合模型预测精度比单一模型预测精度高。     

基于非线性时间序列的预测模型检验与优化的研究

 模型的适用性检验和参数优化是系统建模的最关键环节,对于预测模型的适用性检验,常采用残差方差图、最小信息准则和AIC准则等方法,存在计算量大、准确性低、模型不唯一等缺点.本文给出采用自相关系数和偏自相关系数的拖尾先对ARIMA模型检验,再对其进行F适用性检验,克服了由于观测样本的长度是有限的,偏相关的估计存在误差,拖尾时不能为ARMA定阶的缺陷,并采用具有超线性收敛性等诸多优点的变尺度法对模型参数进行了优化,得到了较为精确的、单一AIRMA 模型,该方法可应用于网络流量模型的适用性检验和模型优化,为网络流量的预测、异常检测和服务器负载预测的应用奠定了坚实的基础.     

移动网无线信号变化预测研究

鉴于移动通信网规模日益扩大,网络运营状态不易即时掌控的现状,提出了一种用于观测移动网无线信号实时变化的新型监控系统。在该系统长期运行而获得的采样数据基础上,对某处小区的无线信号变化特性进行了研究。运用SPSS统计工具和时间序列中的Box—Jenkins的建模方法,分别建立了AR、MA、ARMA、ARIMA模型对实测数据进行了分析和预测,然后对不同模型的预测结果进行了误差分析,结果表明ARIMA（1,1,1）模型准确性最高,误差最小,能对短期内的无线信号变化趋势进行预测。     

含置信区间的ARIMA与GARCH混合式电价预测方法研究

通过对已有的各种电价预测方法的深入研究,提出一种带置信区间的混合式电价预测方法.比较了自相关函数和偏自相关函数判断平稳性及残差方差图定阶法,认为游程检验判断电价序列平稳性方法和AIC准则模型定阶法可避免预测过程中的主观因素,提高预测精度.通过引入电价误差序列异方差判断,结合ARIMA和GARCH,提出含置信区间的电价预测方法,极大地克服了单点预测缺点,增加算法的灵活性,使得市场竞价者可以根据自己对电价预测精度的期望选择电价波动的范围.采用PJM市场数据,验证了算法的有效性.     

Application of neural networks in forecasting engine systems reliability

This paper presents a comparative study of the predictive performances of neural network time series models for forecasting failures and reliability in engine systems. Traditionally, failure data analysis requires specifications of parametric failure distributions and justifications of certain assumptions, which are at times difficult to validate. On the other hand, the time series modeling technique using neural networks provides a promising alternative. Neural network modeling via feed-forward multilayer perceptron (MLP) suffers from local minima problems and long computation time. The radial basis function (RBF) neural network architecture is found to be a viable alternative due to its shorter training time. Illustrative examples using reliability testing and field data showed that the proposed model results in comparable or better predictive performance than traditional MLP model and the linear benchmark based on Box–Jenkins autoregressive-integrated-moving average (ARIMA) models. The effects of input window size and hidden layer nodes are further investigated. Appropriate design topologies can be determined via sensitivity analysis.     

基于ARIMA模型自动预测我国保险行业保费收入的应用和实践

本文采用ARJMA模型以我国财产险的保费收入的预测为例,对数据预处理、模型的选择、识别、估计、检验等过程进行了详细的分析,得出了预测结果报告,并描述了对我国产寿险的保费收入实现自动预测并得到结果报表的思路和方法。     

自回归求和滑动平均方法用于间歇过程变量在线预测

间歇过程变量的在线预测是一种重要的生产过程质量控制手段。实现间歇过程变量的在线预报需要对过程以往的批次数据建立预测模型,即需挖掘批次间和批次内的数据信息。针对间歇过程数据不同批次不等长、数据长度短、非线性等特点,采用数据重构——自回归求和滑动平均方法建立其在线预测模型：将收集到的间歇过程变量以批次为单位进行数据平滑;对这些批次数据按照随机的顺序首尾相接,组成长数据集;对于批次连接处数据跳跃的情况,采用后面所有批次数据减去上一批次的最后一个值,以实现数据的平滑;采用自回归求和滑动平均方法建立数据模型,并用于间歇蒸馏温度的在线预报。采用该方法建立的4步预测模型对某间歇蒸馏过程上升气温度的预测均方差较小,符合生产现场的预测要求。     

维修时点预测的动态车间调度问题

针对制造业中生产计划的不确定问题,提出一种维修时点预测与自适应的遗传模拟退火算法相结合的优化调度方法。该方法首先利用差分自回归移动平均模型预测设备未来的故障率,然后借助电气设备的威布尔（Weibull）分布模型逆向求出设备未来故障发生时刻,最后将此作为约束条件,利用自适应的遗传模拟退火算法解决传统的生产调度问题。结合工厂实际情况,主要分析了设备有无维修的随机调度问题,以最小化最大完工时间为目标,获取每一个任务的调度计划以及每一台设备的维修时点,确定出最佳调度方案。实验表明自适应的遗传模拟退火算法的性能较好。在河北某工厂的生产车间中,设备在运行调度方法后三个月的平均故障率比运行前相对降低了3.46%。