Times Series Prediction to Basis of a Neural Network Conceived by a Real Genetic Algorithm

Neural network and genetic algorithms are complementary technologies in the design of adaptive intelligent system. Neural network learns from scratch by adjusting the interconnections betweens layers. Genetic algorithms are a popular computing framework that uses principals from natural population genetics to evolve solutions to problems. Various forecasting methods have been developed on the basis of neural network, but accuracy has been matter of concern in these forecasts. In neural network methods forecasted values depend to the choose of neural predictor structure, the number of the input, the lag. To remedy to these problem, in this paper, the authors are investing the applicability of an automatic design of a neural predictor realized by real Genetic Algorithms to predict the future value of a time series. The prediction method is tested by using meteorology time series that are daily and weekly mean temperatures in Melbourne, Australia, 1980-1990.

An application of a genetic algorithm in co-optimization of geological CO_(2) storage based on artificial neural networks

Global warming,driven by human-induced disruptions to the natural carbon dioxide(CO_(2))cycle,is a pressing concern.To mitigate this,carbon capture and storage has emerged as a key strategy that enables the continued use of fossil fuels while transitioning to cleaner energy sources.Deep saline aquifers are of particular interest due to their substantial CO_(2) storage potential,often located near fossil fuel reservoirs.In this study,a deep saline aquifer model with a saline water production well was constructed to develop the optimization workflow.Due to the time-consuming nature of each realization of the numerical simulation,we introduce a sur-rogate aquifer model derived from extracted data.The novelty of our work lies in the pioneering of simultaneous optimization using machine learning within an integrated framework.Unlike previous studies,which typically focused on single-parameter optimiza-tion,our research addresses this gap by performing multi-objective optimization for CO_(2) storage and breakthrough time in deep sa-line aquifers using a data-driven model.Our methodology encompasses preprocessing and feature selection,identifying eight pivotal parameters.Evaluation metrics include root mean square error(RMSE),mean absolute percentage error(MAPE)and R^(2).In predicting CO_(2) storage values,RMSE,MAPE and R^(2)in test data were 2.07%,1.52% and 0.99,respectively,while in blind data,they were 2.5%,2.05% and 0.99.For the CO_(2) breakthrough time,RMSE,MAPE and R^(2) in the test data were 2.1%,1.77% and 0.93,while in the blind data they were 2.8%,2.23% and 0.92,respectively.In addressing the substantial computational demands and time-consuming nature of coup-ling a numerical simulator with an optimization algorithm,we have adopted a strategy in which the trained artificial neural network is seamlessly integrated with a multi-objective genetic algorithm.Within this framework,we conducted 5000 comprehensive experi-ments to rigorously validate the development of the Pareto front,highlighting the depth of our computational approach.The findings of the study promise insights into the interplay between CO_(2) breakthrough time and storage in aquifer-based carbon capture and storage processes within an integrated framework based on data-driven coupled multi-objective optimization.

A Hybrid Neural Network-based Approach for Forecasting Water Demand

Water is a vital resource.It supports a multitude of industries,civilizations,and agriculture.However,climatic conditions impact water availability,particularly in desert areas where the temperature is high,and rain is scarce.Therefore,it is crucial to forecast water demand to provide it to sectors either on regular or emergency days.The study aims to develop an accurate model to forecast daily water demand under the impact of climatic conditions.This forecasting is known as a multivariate time series because it uses both the historical data of water demand and climatic conditions to forecast the future.Focusing on the collected data of Jeddah city,Saudi Arabia in the period between 2004 and 2018,we develop a hybrid approach that uses Artificial Neural Networks(ANN)for forecasting and Particle Swarm Optimization algorithm(PSO)for tuning ANNs’hyperparameters.Based on the Root Mean Square Error(RMSE)metric,results show that the(PSO-ANN)is an accurate model for multivariate time series forecasting.Also,the first day is the most difficult day for prediction(highest error rate),while the second day is the easiest to predict(lowest error rate).Finally,correlation analysis shows that the dew point is the most climatic factor affecting water demand.

基于改进小波神经网络的实时系统任务流量预测方法

针对当前航空装备实时系统对非周期实时任务无法预知难以实现可靠调度的困难,开展对航空装备实时系统非周期任务流量预测方法的研究。以小波神经网络为基础结合航空装备实时系统的特性建立任务流量预测模型,并提出利用人工鱼群算法对小波预测模型关键参数进行优化,避免陷入局部最优解,最终构建一种人工鱼群算法改进的小波神经网络任务流量预测系统。利用提出的预测模型开展实时任务流量预测对比仿真实验,实验结果表明,建立的基于改进小波神经网络的实时系统任务流量预测系统对非周期实时任务具有较高的预测精度,预测效果优于原始小波神经网络模型及T-S模糊神经网络模型。

Modeling the Nigerian Bonny Light Crude Oil Price: The Power of Fuzzy Time Series

Several authors have used different classical statistical models to fit the Nigerian Bonny Light crude oil price but the application of machine learning models and Fuzzy Time Series model on the crude oil price has been grossly understudied. Therefore, in this study, a classical statistical model—Autoregressive Integrated Moving Average (ARIMA), two machine learning models—Artificial Neural Network (ANN) and Random Forest (RF) and Fuzzy Time Series (FTS) Model were compared in modeling the Nigerian Bonny Light crude oil price data for the periods from January, 2006 to December, 2020. The monthly secondary data were collected from the Nigerian National Petroleum Corporation (NNPC) and Reuters website and divided into train (70%) and test (30%) sets. The train set was used in building the models and the models were validated using the test set. The performance measures used for the comparison include: The modified Diebold-Mariano test, the Root Mean Square Error (RMSE), the Mean Absolute Percentage Error (MAPE) and Nash-Sutcliffe Efficiency (NSE) values. Based on the performance measures, ANN (4, 1, 1) and RF performed better than ARIMA (1, 1, 0) model but FTS model using Chen’s algorithm outperformed every other model. The results recommend the use of FTS model for forecasting future values of the Nigerian Bonny Light Crude oil. However, a hybrid model of ARIMA-ANN or ARIMA-RF should be built and compared with Chen’s algorithm FTS model for the same data set to further verify the power of FTS model using Chen’s algorithm.

基于近红外光谱技术的远安黄茶品质快速无损检测方法

应用近红外光谱(near infrared spectroscopy,NIRS)技术结合多种算法对远安黄茶品质开展快速无损评价。首先通过扫描获得远安黄茶90个样品的近红外光谱,再利用11种不同方法对光谱进行预处理,剔除部分噪声信息,然后应用反向区间偏最小二乘法(backward interval partial least squares,Bi-PLS)筛选反映样品品质的特征光谱区间,应用遗传算法(genetic algorithm,GA)精准提取特征光谱波长,建立了5种黄茶品质预测模型,最后对光谱官能团信息进行解析。结果表明,最佳光谱预处理方法为多元散射校正,Bi-PLS筛选出的特征光谱区间主要为9003.2~7497.9 cm^(-1)、6101.7~5449.8 cm^(-1)和4601.3~4246.5 cm^(-1),GA筛选出75个特征光谱波长,建立的Bi-GA-PLS组合模型具有最佳的稳健性,可准确地预测远安黄茶样品外部品质分数(R2=0.951,RMSEP=1.57,RPD=5.27),初步实现了远安黄茶品质的快速、准确预测。光谱信息解析结果显示,45个光谱波长反映—CHx、C=O和—NHx官能团信息,代表单糖、咖啡碱、茶氨酸和游离蛋白质等内含成分物质,30个光谱波长反映O—H、酰胺键以及C—H和C—C伸缩的组合频信息,代表木质素、淀粉、纤维素等多糖内含成分物质。

暂态稳定仿真的综合人工智能方法

研究了利用综合人工智能进行暂态稳定仿真的方法:利用神经网络训练动态元件的输入输出关系,同时,利用遗传算法优化神经网络的结构和加快神经网络的训练速度。算例结果表明,利用遗传算法和神经网络相结合的综合人工智能方法暂态稳定仿真比传统的时域仿真计算速度快,并能保持较高的精度,可以有效地提高暂态稳定仿真的实时性。

自组织递归区间二型模糊神经网络在动态时变系统辨识中的应用

针对动态时变系统辨识过程中存在噪声干扰的问题,本文将区间二型模糊集结合到递归神经网络中,提出了自组织递归区间二型模糊神经网络以增强动态时变系统的抗噪能力。该自组织递归区间二型模糊神经网络由前件和后件两部分构成:前件为区间二型模糊集模型,用于将每个规则的激活强度反馈到自身构成内反馈回路,其参数学习采用梯度下降算法;后件为带有区间权值的Takagi-Sugeno-Kang(TSK)模型,其参数学习采用有序规则卡尔曼滤波算法,且网络初始规则数为零。所有规则均通过结构学习和前后件参数同时在线学习来产生,其网络结构学习采用的是在线区间二型模糊群集。为验证提出的神经网络的优越性,将其应用到单输入单输出动态时变系统的辨识中。实验结果表明,相对于前馈一型/二型模糊神经网络、递归一型模糊神经网络,该神经网络的辨识能力强,即使在存在白噪声的条件下,也能减小测试及训练误差。

模糊神经网络在时间序列预测中的应用

文中提出了将模糊聚类与梯度算法相结合的一种改进的训练模糊神经网络的混合型算法．模拟结果表明，模糊神经网络可以成功地用于时间序列的预测，模糊神经网络的训练速度与模拟精度都优于传统多层ＢＰ网络．

基于改进AFSA的路面不平度时域估测

为提高根据车辆动力学响应建立路面不平度时域模型的精度,对RBF神经网络的设计、输入神经网络的动力学响应参数和汽车车体测量点的位置进行了研究。基于拉格朗日第2方程建立了车身任意位置的5自由度振动模型,以滤波白噪声法建立的路面时域激励为车辆激励的输入和神经网络的理想输出,采用改进的人工鱼群算法(AFSA),针对车身测量点距质心的距离、待测量的动力学响应参数的类型和RBF神经网络的扩展系数建立了优化分析模型。提出了2种需测量的车辆动力学响应参数方案,以及各方案下车身测量点的具体位置。研究结果表明,2种方案的路面不平度时域激励估测精度均高于0.99。

生菜叶中磷含量的光谱定量分析

为快速、准确检测生菜叶内的磷含量,提出了应用光谱技术结合化学计量法无损检测生菜叶内磷含量的方法。通过获取不同施磷量下生菜叶片于波长350~2 500 nm处的反射光谱,对光谱数据进行5点平滑和一阶导数变换后,利用联合区间偏最小二乘算法(si PLS)提取了与生菜叶磷元素相关的4个特征波段,即950~1 070 nm,1 430~1 549 nm,1 906~2 025 nm和2 144~2 263 nm。进一步利用连续投影算法(SPA)对全光谱波段和4个特征波段进行特征波长提取,分别筛选出变量63个和25个。分别对4个特征波段、63个和25个特征波长进行主成分降维,当主成分数分别为7、5和4时,隐含层神经元数分别为7、5和3时,建立了si PLS+BPANN,SPA+BPANN,si PLS+SPA+BPANN生菜叶磷含量检测模型。研究结果表明:si PLS+SPA+BPANN模型的预测结果优于其他模型,验证集相关系数为0.911,验证均方根误差为479 mg/kg。

基于多尺度熵的滚动轴承Elman神经网络故障诊断方法

针对滚动轴承故障振动信号具有跨尺度复杂性的特点,提出了一种新的基于多尺度熵(multiscale entropy,MSE)和反馈式Elman神经网络的滚动轴承故障诊断方法。该方法利用MSE对滚动轴承不同健康状态下的振动信号进行故障特征提取,并将其作为Elman神经网络的输入,利用Elman神经网络自动识别轴承所属的故障类型及故障程度。实验数据包括不同故障类型和不同故障程度样本,结果表明提出的方法能有效地实现滚动轴承故障类型以及程度的智能诊断,效果优于前馈式概率神经网络(Probabilistic neural network,PNN),并具有较低的虚警率和漏警率。

基于遗传神经网络的渗流实时预报方法研究

考虑渗流的主要影响因子,运用遗传神经网络建立了大坝渗流实时预报模型。该模型具有再学习能力,在应用过程中,可以用新的观测资料对模型不断地进行学习训练,且随着样本的积累,模型预报精度不断提高,预报速度很快,因而完全满足实时预报的需要。利用该模型对丰满大坝横向扬压力进行了建模预报,预报值与实测值基本吻合,预测精度高,证明该方法用于大坝渗流实时预报的有效性和实用性。

基于改进型BP算法的外债风险指标预测

利用人工神经网络进行时间序列预测是一种较新的方法 ,它具有不需建立复杂的数学模型以及非线性映射能力强等优点 .采用动量法和学习率自适应调整的改进型 BP算法对外债风险的各项指标进行了非线性时间序列的预测 。

液体火箭发动机启动过程实时在线故障检测算法

利用神经网络技术实现了液体火箭发动机启动过程的非线性辩识；提出并实现了一种基于辩识误差检验的故障检测策略。经大量实际发动机热试车数据验证表明，所提出的检测算法十分有效。由于算法所利用的监测参数均系实际发动机地面试车中所测量的参数，且检测算法在线工作时计算量十分小，因而所提出并实现的检测算法可以直接应用于工程实际。

基于混沌搜索的人工蜂群优化神经网络交通流预测方法

为了提高长时交通流的预测精度,提出一种改进的人工蜂群优化BP神经网络分时段预测交通流的方法;利用Tent混沌映射采蜜蜂放弃的新解,实现具有混沌搜索策略的人工蜂群算法,然后优化BP神经网络的权值和阈值,最终训练BP神经网络以求得最优值;利用该预测方法对合肥市黄天路全天的交通流分时段预测,实现了对长时交通流的准确预测,与传统的人工蜂群优化BP神经网络预测对比,能有效改善预测精度,降低预测误差。

递阶遗传粒子群算法在神经网络设计中的应用

将递阶遗传粒子群算法(HGAPSO)应用于神经网络设计,可以在对网络拓扑结构优化的同时对连接权重进行求解。该算法结合了遗传算法在解决离散问题和粒子群算法在解决连续问题上的优势,并利用BP算法沿误差最速下降的能力对连接权重进一步学习,达到全局最优和快速搜索的有机结合。通过对混沌时序信号的预测,表明递阶遗传粒子群算法在较大程度上提高了神经网络的学习性能和泛化能力。

智能预测在三峡永久船闸中隔墩时效变形分析中的应用

人工神经网络通过调整网络中的连接权值、网络结构,实现输入输出非线性映射关系,对三峡永久船闸中隔墩的变形进行预测。具有全局最优化遗传算法可使连接权值、网络结构得到最优解。将遗传算法与神经网络结合起来进行变形演化特征识别,对三峡永久船闸中隔墩时效变形进行智能预测,应用结果表明该方法具有较高的预测精度。

一种优化的神经网络数字预失真方法

提出一种基于遗传算法和低阶广义记忆多项式实值神经网络的射频功率放大器数字预失真方法。该方法将遗传算法优化的低阶广义记忆多项式模型与神经网络模型进行级联来增强校正模型与功放失真的匹配程度。它不仅可以提升模型的校正能力,同时可以加快网络的收敛速度。采用60MHz的三载波LTE信号进行实验,通过与实值延时线神经网络模型对比,在收敛速度上有显著提升,同时在邻道功率泄露ACLR指标上有6 d B左右改善。

基于ANN-PSO算法的pH依赖-时滞型地锦草结肠靶向给药微丸制备工艺优化

目的应用人工神经网络结合粒子群优化法(ANN-PSO)筛选地锦草结肠靶向给药微丸制备工艺参数。方法以pH依赖-时滞型地锦草结肠靶向给药微丸的制备为研究模型,薄膜包衣增重、增塑剂与成膜材料用量比为变量,微丸体外释放总评归一值为因变量,应用BP(back-propagation,反向传播)人工神经网络建模,并结合粒子群优化法筛选微丸的处方工艺参数。结果按优化的处方工艺参数制备的地锦草结肠靶向给药微丸,在体外实验中可满足结肠定位释放的要求。结论人工神经网络建模与粒子群法寻优相结合,为解决药物制剂工艺涉及的多维复杂非线性系统的优化问题提供了有效的途径。