基于MPSO-WLS-SVM的矿井瓦斯涌出量预测模型研究

为有效预防瓦斯灾害,以预测矿井瓦斯涌出量为研究目的,提出经改进的粒子群算法(MPSO)优化的加权最小二乘支持向量机(WLS-SVM),并用其预测非线性动态瓦斯涌出量。算法通过对WLS-SVM的正则化参数C和高斯核参数σ寻优,建立基于MPSO优化的WLS-SVM的瓦斯涌出量预测模型,并利用某矿井监测到的各项历史数据进行实例分析。试验结果表明:该预测模型预测的最大相对误差为5.99%,最小相对误差为0.43%,平均相对误差为2.95%,较其他预测模型有更强的泛化能力和更高的预测精度。     

Action-Oriented Support for Occupational Safety and Health Programs in Some Developing Countries in Asia

Action-oriented support programs have been increasingly playing vital roles in promoting safety and health in developing countries in Asia. This paper studied achievements of 3 action-oriented support programs: the WISE program for small enterprises, the WIND program for farmers, and the POSITIVE program for workers and trade unions. Special attention was paid to how the programs have strengthened local efforts for sustainable actions in safety and health improvements. The results showed that there were significant achievements in action-oriented support programs in the region, including a large number of improvement examples, integration into government policies, and network support through employers’ and workers’ organizations. Participatory, action-oriented training tools such as action-checklists, local good example photos, and group work methods played key roles in the effective implementation of the programs. It was of note that there were a number of local efforts to extend the coverage of action-oriented support even to hard-to-reach workers such as home-based workers, rural workers, and ethnic minorities. The efforts included the equal participation in the training by female and male farmers, shortened and weekend training programs, photo sheets showing local good examples, and reasonable fee collection for better sustainability. In conclusion, action-oriented support programs provided local people with concrete means to promote safety and health improvements. The successful programs commonly focused on local initiatives and were built on local wisdom and resources.     

The effects of perceived organizational support,perceived supervisor support and perceived co-worker support on safety and health compliance

Although knowledge is cumulating, very little is known about the effects of various sources of support on safety and health compliance. This study goes beyond previous research by investigating the relationships among perceived support from organizations, supervisors and co-workers, and employees’ safety and health compliance behaviour at chemical and petroleum process plants. The results of this study show that the support from organizations, supervisors and co-workers was significantly related to employees’ safety and health compliance. Also, the findings reveal that perceived supervisor support has the strongest influence in ensuring employees’ safety and health compliance behaviour.     

When are overqualified employees creative? It depends on contextual factors

The research on perceived overqualification has mainly examined its negative consequences. Defined, employees who feel overqualified have surplus talent and thus can be excellent workers if managed properly; yet, empirical evidence in this domain is lacking. Building on person–environment fit theory, this research explored whether, when, and how employees who feel overqualified can engage in creative performance. The results of a multi‐source field study (N = 113 employees and 19 supervisors) supported theoretical predictions. Perceived overqualification was related positively to supervisor‐rated creative performance when these workers felt supported and appreciated and successfully negotiated developmental idiosyncratic deals. Opportunities to mentor others had an impact on the relationship between perceived overqualification and supervisor‐rated creativity, although the simple slopes were non‐significant. This study is novel in that it unpacked actionable steps that organizations can utilize to motivate this large segment of workforce to use their surplus qualifications constructively by, for example, engaging in creative performance. Copyright © 2015 John Wiley & Sons, Ltd.     

支持向量机在边坡稳定性预测中的应用

针对离子型稀土原地浸矿边坡稳定性问题的小样本、非线性等特点,采用一种边坡稳定性预测模型——支持向量机(SVM)模型,分别使用网格参数寻优、遗传算法(GA)参数寻优、粒子群(PSO)寻优算法来计算SVM模型的参数,分析比较3种算法得到的参数,确定最适合离子型稀土原地浸矿边坡稳定性预测的SVM回归模型的参数寻优算法。     

基于集对论和主客观综合赋权法的装备保障能力评估研究

目的提出一种基于集对论和主客观赋权法的装备保障能力评估方法。方法针对装备保障能力评估中存在的诸多不确定问题,构建装备保障能力评估指标体系,基于集对论思想建立装备保障能力集对评估模型。结果通过构造一个双目标决策问题,进行主客观综合赋权确定指标权重,使得评估指标权重向量同时反映了主观程度和客观程度。结论该方法通过联系度量化了评估指标的确定程度,提高了评估结果的可信性,对解决装备保障能力评估中不确定性问题具有很强的指导作用。     

基于PCA-AHPSO-SVR的煤层瓦斯含量预测研究

为了提高煤层瓦斯含量预测的准确性和科学性,通过主成分分析方法对影响煤层瓦斯含量的7个因素进行特征提取,消除影响因素之间的相关性,减少维度;用支持向量回归机对提取的因素进行训练,并用改进的自适应混合粒子群算法对SVR的参数进行优化,提出PCA-AHPSO-SVR模型;与PCA-PSO-SVR,PSO-SVR这2个模型在相同环境下进行30次运行比较。研究结果表明:研究提出的PCA-AHPSO-SVR模型较其他2种模型平均准确率分别提高5.51%和9.32%,稳定性更佳,可满足工程实际需求。     

高应力大变形软岩巷道“三壳”围岩控制机理及应用

为解决土城煤矿14运输上山软岩巷道变形量大、锚杆索失效严重等技术难题,通过现场调研、室内试验、理论分析、数值模拟及工业性试验等方法,揭示了围岩变形特征以及巷道失稳破坏机理,提出了“三壳”围岩控制理论。基于以上研究,设计了“锚杆锚索+灌浆+钢管混凝土支架”的复合支护方案,建立了基于“三壳”围岩控制理论的“三壳”支护结构体力学模型,计算出设计方案的极限承载能力为2.54 MPa,随后采用 FLAC3D数值模拟软件对设计方案进行模拟分析,验证了方案合理性。最后,该复合支护得到成功运用,现场监测结果表明:巷道顶底板以及两帮变形量均低于100 mm,巷道未发生明显变形,支护效果良好。     

基于属性约简的绝缘节破损预测

准确的绝缘节破损预测能够保证铁路运输安全和经济效益。支持向量机算法能够处理轨道电路测试数据,对其进行分类,预测可能存在隐患的绝缘节,但支持向量机预测模型的原始样本多有冗余,基于此,提出了一种基于粗糙集和支持向量机的绝缘节破损预测模型。通过改进主分量启发式属性约简算法,降低样本维数,同时选用模拟退火算法完成SVM自动参数选优。实例分析和仿真结果表明,与单一支持向量机算法相比,属性约简后的粗糙集-支持向量机算法提高了分类器的分类性能,与采用网格搜索技术的SVM预测方法相比,模拟退火算法有效提高了SVM的预测精度。     

支持向量机对多环芳烃毒性的定量构效预测

为了实现多环芳烃(PAHs)毒性的有效预测,提出应用定量构效技术对多环芳烃的空气-正辛醇分配系数(KOA)和致癌性进行预测。应用分子描述符和试验值确立构效关系,采用支持向量机算法(SVM)和人工神经网络算法(ANN)分别建立了PAHs的KOA回归预测模型和致癌性分类预测模型。利用网格划分(GS)、遗传算法(GA)、粒子群算法(PSO)对SVM进行参数寻优。应用均方误差(MSE)、拟合决定系数R2和分类准确率(Accuracy)分别对模型进行了验证与评价。结果表明,最佳回归预测模型GS-SVR的MSE为0.059 7,R2为0.913 0;最佳分类预测模型GA-SVC的Accuracy为95%。研究表明:应用SVM所建两种模型的稳定性和预测能力都优于应用ANN建立的模型;参数优化后模型的稳定性和预测能力得到了提高。