基于关系图卷积网络的代码搜索方法

代码搜索是当下自然语言处理和软件工程交叉领域的一个重要分支. 开发高效的代码搜索算法能够显著提高代码重用的能力, 从而有效提高软件开发人员的工作效率. 代码搜索任务是以描述代码片段功能的自然语言作为输入, 在海量代码库中搜索得到相关代码片段的过程. 基于序列模型的代码搜索方法DeepCS虽然取得了很好的效果, 但这种方法不能捕捉代码的深层语义. 基于图嵌入的代码搜索方法GraphSearchNet能缓解这个问题, 但没有对代码与文本进行细粒度匹配, 也忽视了代码图和文本图的全局关系. 为了解决以上局限性, 提出基于关系图卷积网络的代码搜索方法, 对构建的文本图和代码图编码, 从节点层面对文本查询和代码片段进行细粒度匹配, 并应用神经张量网络捕捉它们的全局关系. 在两个公开数据集上的实验结果表明, 所提方法比先进的基线模型DeepCS和GraphSearchNet搜索精度更高.     

改进乌鸦搜索算法在邮件集散中心航空运力调度中的应用

邮件集散中心航空运力调度涉及固定运力和备选运力两种调度对象,本文在航空运力资源充足的前提下,建立了以最小化运输成本为目标的优化模型,研究了一种改进的乌鸦搜索算法求解方法.首先根据问题的数学模型,引入惩罚函数法将部分约束转化为惩罚项,与目标函数共同构成适应度函数;然后引入Logistic混沌映射提高初始种群的多样性;根据问题的特点,提出了基于个体最优追随机制和正余弦算法的位置更新策略,并引入交叉变异机制以丰富搜索过程中种群的多样性.通过大量算例实验分析,证明了该算法的有效性与优越性.     

联合变换图像识别中功率谱高斯掩膜处理方法

本文提出了一种基于在光学频谱面上进行光学高斯高通滤波的实时处理方法.此方法将一个高斯掩膜覆盖在位于光学频谱面上的CCD中心位置,从而明显降低了相关图像中自相关峰的高度和宽度.通过比较不同的参数效果,确定了一种用于实验的高斯函数滤波器.通过系列实验研究发现,本方法可以明显提高联合变换相关器的效果和速度.     

湖南省峰谷分时电价分析及优化策略研究

针对湖南省电网执行峰谷分时电价存在的问题,结合湖南省实际情况,提出完善峰谷分时电价的策略.以规划综合大数据平台为基础,基于负荷需求弹性分析,综合考虑供电成本、用户用电成本、负荷优化目标等条件,通过研究电价价差调整原则,制定出不同时段的电价,根据电价的经济信号,引导各类用户合理用电,优化负荷特性,提升湖南电力系统运行效率...     

基于高斯插值提高雷达测距精度的研究

本文结合脉冲法雷达测距原理,分析了距离量化误差产生的原因.针对LFM信号加入高斯白噪声作脉冲压缩得到的脉压波形,结合先验信息,将脉压结果主瓣内的波形近似为高斯分布,利用脉压结果中采样最大值点和次大值点的位置信息和幅度信息作高斯插值处理,预估脉压波形真实峰值时间.仿真结果表明,通过高斯插值处理,距离量化误差显著减小,雷达...     

四川地区大规模水电集群停机避峰特征分析及管理系统设计

近年来,随着四川地区水电站数量和装机规模的急剧增加,大规模水电集群同时采取停机避峰措施对电网安全运行的影响日益突出。文中结合四川地区水电站运行、水电调度管理和电网运行实际情况,明确了引发水电站停机避峰的主要因素为流量、泥沙和杂质,分析了大规模集群径流式水电站停机避峰对电网调度的影响;从时间和空间(流域)2个维度研究了停机避峰发生的分布特征,提出了以流量和水质作为预测水电站是否发生停机避峰的可行判断方法;从预测方法、实时监测和系统架构等方面介绍大规模水电集群停机避峰管理系统设计思路。所设计的管理系统在四川电网的实际应用表明,该系统可以提高电网调度机构对实时停机避峰掌控能力,有助于辅助电网汛期安全稳定运行,可以为区域电网和水电站厂停机避峰管理提供参考。     

Exploratory product search using top-k join queries

Given a relation that contains main products and a set of relations corresponding to accessory products that can be combined with a main product, the Exploratory Top-k Join query retrieves the k best combinations of main and accessory products based on user preferences. As a result, the user is presented with a set of k combinations of distinct main products, where a main product is combined with accessory products only if the combination has a better score than the single main product. We model this problem as a rank-join problem, where each combination is represented by a tuple from the main relation and a set of tuples from (some of) the accessory relations. The nature of the problem is challenging because the inclusion of accessory products is not predefined by the user, but instead all potential combinations (joins) are explored during query processing in order to identify the highest scoring combinations. Existing approaches cannot be directly applied to this problem, as they are designed for joining a predefined set of relations. In this paper, we present algorithms for processing exploratory top-k joins that adopt the pull-bound framework for rank-join processing. We introduce a novel algorithm (XRJN) which employs a more efficient bounding scheme and allows earlier termination of query processing. We also provide theoretical guarantees on the performance of this algorithm, by proving that XRJN is instance-optimal. In addition, we consider a pulling strategy that boosts the performance of query processing even further. Finally, we conduct a detailed experimental study that demonstrates the efficiency of the proposed algorithms in various setups.     

Maximum vertex occupation time and inert fugitive: Recontamination does help

Given a simple graph G, we consider the node search problem with inert fugitive. We are interested in minimizing the maximum vertex occupation time, i.e. the maximum number of steps in which a vertex is occupied by a searcher during a search of G. We prove that a search program which does not allow a recontamination may not find an optimal solution to this problem. Moreover, the difference between the minimum maximum vertex occupation time computed by a monotone search program and a program without such a restriction may be arbitrarily large.     

基于样条插值的爆破振动时域波形预测方法

针对现有的爆破振动预测方法存在工程现场适用性差,不易于应用,无法准确预测完整的爆破振动信号在爆破时间段内的振动历程等问题,提出了一种基于样条插值的爆破振动预测方法。通过爆破中质点峰值速度、炸药量与爆心距的关系式,建立监测波形与预测波形的关系,对实测数据波形时域部分进行三次样条插值得到插值函数和预测目标点波形。通过实验和工程实践对该方法进行验证,并与传统的爆破振动预测方法进行质点峰值速度比对,结果表明:该爆破振动预测方法能够完整且较好地预测振动信号在爆破时间段内的振动历程,预测效果优于传统的爆破振动预测方法。     

Effects of magnification methods and magnifier shapes on visual inspection

This study investigated the effectiveness of the novel differential linear and differential nonlinear magnification methods, and the effects of magnifier shape and magnification power on visual search performance. With the differential magnification, objects that were more peripheral were scaled at a level higher than those in the vicinity of fixation. It was hypothesized that 1) owing to the ineffectiveness of scaling stimulus in the vicinity of fixation, the differential magnification methods would enhance magnification effectiveness and thereby improve inspection performance, and 2) because an elliptical lens shape resembles the shape of human visual field, for the same area of magnified view, an elliptical magnifier would be more effective than the circular one. In this study, objective performance measures of speed and accuracy and subjective workload evaluation using the NASA Task Load Index paradigm were used. The results showed that 4× magnification increased the overall inspection time as compared with 3× magnification, suggesting that a trade-off might have occurred between magnification and field size. The two hypotheses were not confirmed; however, a number of questions arose which need further investigation. Discussions of the concerned questions were made on i) the unfamiliarity of subjects with the new differential magnification methods, ii) perceptual discomfort due to image duplication, discontinuity, and distortion resulting from stimulus scaling, iii) the movement of fixation points, and iv) the competition for attention between foveal and peripheral magnified objects. This study prompts the authors to consider a training programme on the differential linear and nonlinear magnification methods in future studies. With proper instructions regarding the nature of lens and magnified images, subjects' performance with the novel differential magnification methods may be improved.