基于NR-Transformer的集群作业运行时间预测

高性能集群的作业调度通常使用作业调度系统来实现，准确填写作业运行时间能在很大程度上提升作业调度效率。现有的研究通常使用机器学习的预测方式，在预测精度和实用性上还存在一定的提升空间。为了进一步提高集群作业运行时间预测的准确率，考虑先对集群作业日志进行聚类，将作业类别信息添加到作业特征中，再使用基于注意力机制的NR-Transformer网络对作业日志数据建模和预测。在数据处理上，根据与预测目标的相关性、特征的完整性和数据的有效性，从历史日志数据集中筛选出7维特征，并按作业运行时间的长度将其划分为多个作业集，再对各作业集分别进行训练和预测。实验结果表明，相比于传统机器学习和BP神经网络，时序神经网络结构有更好的预测性能，其中NR-Transformer在各作业集上都有较好的性能。     

不同粒度纳米碳酸钙的可控制备

不同粒度的纳米碳酸钙具有不同的特性和应用范围,目前不同粒度纳米碳酸钙的可控制备方法还未见报道。以氯化钙、碳酸铵为原料,以乙醇、柠檬酸和焦磷酸钠为分散剂,采用复分解法研究了不同制备条件对纳米碳酸钙粒径的影响规律,在此基础上制备出平均粒径为17~71 nm的碳酸钙,并对其进行了表征。实验结果表明,所制备的纳米碳酸钙均为球霰石型,纯度很高,而且形貌近似球形。制备条件对纳米碳酸钙的粒径有显著的影响：随着反应温度升高,纳米碳酸钙的粒径减小;随着氯化钙滴加时间的增加,纳米碳酸钙的粒径增大;随着反应物浓度的提高,纳米碳酸钙的粒径先减小后增大;当分散剂为柠檬酸和焦磷酸钠的混合溶液时所得纳米碳酸钙的粒径最小,当分散剂为乙醇和焦磷酸钠的混合溶液时所得纳米碳酸钙的粒径最大。利用这些影响规律通过控制制备条件可以实现所需粒径纳米碳酸钙的可控制备。     

不同粒径改性粉煤灰对磷酸根吸附性能的影响

废水排放过量的磷导致水体污染日益严重，将粉煤灰通过化学改性制备成了水化硅酸钙吸附剂，研究了改性吸附剂对磷酸根的吸附效果。利用XRD, SEM及BET比表面积等手段对粒度分级前后的吸附剂进行表征，研究不同粒级吸附剂对磷酸根的吸附性能，并考察其吸附机理。结果表明，不同粒级的吸附剂其化学成分出现了明显的偏析现象，孔隙结构也差异显著。相比其他粒径下的吸附剂颗粒，颗粒粒径在50?75 μm时，吸附剂中钙和硅含量较多，铝、铁和镁含量较低，水化硅酸钙组分含量最高，且伴有含铝的托贝莫来石晶体出现，钙离子的增加使其可以与更多的磷酸根结合形成沉淀。同时此粒径下具有较高的比表面积及孔隙度，疏松多孔的结构为钙离子提供更多活性位点。当使用粒径在50?75 μm的吸附剂吸附磷酸根时，磷的饱和吸附量可达到17.1 mg/g，比未分级的吸附剂高19.58%。     

Job accessibility impacts of intensive and multiple land-use scenarios for the Netherlands’ Randstad Area

This paper examines job accessibility impacts of intensive and multiple land-use scenarios in the Netherlands’ Randstad Area for the period 1996–2030. The scenarios involve a heavy increase in the density and diversity of activities around railway stations. Accessibility impacts are examined using location-based accessibility measures and utility-based accessibility benefit measures within a high-resolution land-use/transport interaction modelling framework, taking into account the regional demand for economic activities and possible market responses. It is concluded that extreme concentrations of economic activities at central urban locations well served by public transport are not to be preferred over less heavy concentrations of economic activities at more locations. Heavy concentrations of new firms may firstly repel existing firms in the neighbourhood and secondly result in decreasing marginal returns of accessibility benefits for public transport users.     

Is activity-based working impacting health,work performance and perceptions? A systematic review

ABSTRACT

Numerous claims have been made about the benefits of activity-based working (ABW) on workers’ health and work performance. Yet, it is unclear if these claims are proven. This systematic review aims to establish whether there is an evidence base for the effects of ABW on health, work performance and perceptions of the work environment. Eight databases were searched in September 2016. Three reviewers independently screened titles and abstracts and assessed the studies and extracted the data. Seventeen studies involving 36,039 participants were included. The study designs varied in rigorousness from qualitative studies to pre–post-trials and in sample size ranging from 12 to 11,799. This review found that ABW has positive merits in the areas of interaction, communication, control of time and space, and satisfaction with the workspace; however, it is unfavourable for concentration and privacy. For physical and mental health, the evidence is equivocal. ABW seems to be a promising concept that can be implemented and promoted based on some benefits for work performance and perceptions of the work environment when it is coupled with appropriate management support and organization. More high-quality research is needed to strengthen the evidence base further and establish its health effects.     

西部弱胶结砂岩循环载荷作用下塑性应变能变化规律

通过不同应力路径下的三轴循环加卸载实验,对中国西部矿区中不同粒径弱胶结砂岩的塑性应变能和塑性变形变化规律进行了研究。实验结果表明,加载路径对砂岩的塑性应变能及塑性变形特征有显著影响;在加载过程中,粗砂,中砂,细砂储存的塑性应变能以及累计速度依次递减,并且塑性应变能随应力增大按指数规律变化;各加卸载周期内储存的塑性应变能随加卸载次数及应力的增大呈“U”字形变化,即初始阶段由于孔隙裂隙压密,会储存较多塑性应变能;弹性阶段变形源于晶格弹性压缩,卸载时可以恢复,储存的塑性应变能较少;屈服阶段由于破裂面滑移也会储存较多塑性应变能;轴向和径向残余应变随应力增大呈指数形式增长,并且径向残余应变增长速度大于轴向;残余应变与塑性应变能变化规律不同,塑性应变能更能真实反映岩石在受载过程中的状态变化。     

墙体内热湿耦合过程的时域递归展开算法

利用时域递归展开算法对墙体内热湿耦合传递方程进行求解。以木板为例,应用该算法进行了热湿耦合传递的分析计算,在时间域和空间域上分别运用递归展开法和控制容积法进行离散,从而得到递归形式的线性方程组,运用MATLAB软件对这一过程进行求解。计算结果与有限差分算法、解析解计算结果以及实验数据吻合良好,表明该算法能够用于求解多孔介质热湿耦合传递模型。时间步长的改变对计算结果影响较小,可通过增加时间步长方法来减少工作量。     

随机工时下柔性加工车间的鲁棒优化调度方法

为了提高柔性加工车间调度方案的可行性、保障生产过程的稳定性,提出一种鲁棒优化调度方法。引入两个不确定参数来描述随机工时的波动程度和约束条件的允许违背程度,提出随机变量服从概率分布时一般线性规划问题的鲁棒优化方法。采用该方法将含随机工时而难以求解的随机型柔性加工车间调度模型转化为确定型鲁棒对等模型。基于该模型,将随机工时融入适应度函数中,结合遗传进化的全局优化和邻域搜索的空间拓展能力研制出鲁棒调度算法,同步实现工件排序和机器分配的双重决策。案例测试表明,所提方法可以在较短计算时间内、以较小性能损失、将近95%的置信度获得当前最优解。