基于遗传算法的湿陷性黄土地基灰土挤密桩优化设计

利用Visual C＋＋语言，定义了一个面向对象的基于实数编码遗传算法类。运用可最大限度保证遗传算法搜索到问题的全局最优解的深度搜索策略，编制了灰土挤密桩的优化设计和造价估算遗传算法程序，并针对兰州某地灰土挤密桩复合地基进行了计算分析，并与原有方案比较，结果表明：两种优化方案在满足承载力和变形的条件下，在降低造价成本方面优势非常明显。     

调运问题中基于栅格模型的快速路径规划方法

针对调运路径规划这一问题,采用栅格模型表示环境地图,通过设定路径搜索方向权重,剔除不必要的搜索区域,提高了搜索效率.仿真结果表明,该算法能有效地提高路径搜索效率,并能搜索到最优路径.     

利用深度学习的硬件计数器复用估计算法

利用深度学习方法,为硬件计数器复用(multiplexing,MPX)提供结果精度更高的估计模型。通过对MPX估计得到的结果与实际采集的真实数据进行相似性分析,证明相同程序多次运行之间得到的硬件计数值是线性相关的。采用神经网络多层感知器(multilayer perceptron,MLP)和双向门控神经网络(bidirectional gated recurrent unit, Bi-GRU)这2种深度学习模型,对MPX数据进行拟合。基于动态时间规整(dynamic time warping, DTW),提出一个全新的评估MPX数据精度的指标DTW-cost。实验结果表明,同时收集15个硬件事件数据时,MLP方法拟合得到的13个高性能计算应用平均准确率比现有使用最广的固定插值法高出10.53%,最多可提升19.8%;而在MLP表现较差的事件上,Bi-GRU方法得到的平均准确率提升了28.8%。     

3D laser scanning strategy based on cascaded deep neural network

A 3D laser scanning strategy based on cascaded deep neural network is proposed for the scanning system converted from 2D Lidar with a pitching motion device. The strategy is aimed at moving target detection and monitoring. Combining the device characteristics, the strategy first proposes a cascaded deep neural network, which inputs 2D point cloud, color image and pitching angle. The outputs are target distance and speed classification. And the cross-entropy loss function of network is modified by using focal loss and uniform distribution to improve the recognition accuracy. Then a pitching range and speed model are proposed to determine pitching motion parameters. Finally, the adaptive scanning is realized by integral separate speed PID. The experimental results show that the accuracies of the improved network target detection box, distance and speed classification are 90.17%, 96.87% and 96.97%, respectively. The average speed error of the improved PID is 0.4239°/s, and the average strategy execution time is 0.1521 s. The range and speed model can effectively reduce the collection of useless information and the deformation of the target point cloud. Conclusively, the experimental of overall scanning strategy show that it can improve target point cloud integrity and density while ensuring the capture of target.     

Discrete search allocation game with false contacts

This paper deals with a two‐person zero‐sum game called a search allocation game, where a searcher and a target participate, taking account of false contacts. The searcher distributes his search effort in a search space in order to detect the target. On the other hand, the target moves to avoid the searcher. As a payoff of the game, we take the cumulative amount of search effort weighted by the target distribution, which can be derived as an approximation of the detection probability of the target. The searcher's strategy is a plan of distributing search effort and the target's is a movement represented by a path or transition probability across the search space. In the search, there are false contacts caused by environmental noises, signal processing noises, or real objects resembling true targets. If they happen, the searcher must take some time for their investigation, which interrupts the search for a while. There have been few researches dealing with search games with false contacts. In this paper, we formulate the game into a mathematical programming problem to obtain its equilibrium point. © 2006 Wiley Periodicals, Inc. Naval Research Logistics, 2007     

A m‐parallel crane scheduling problem with a non‐crossing constraint

In this paper, we study a m‐parallel machine scheduling problem with a non‐crossing constraint motivated by crane scheduling in ports. We decompose the problem to allow time allocations to be determined once crane assignments are known and construct a backtracking search scheme that manipulates domain reduction and pruning strategies. Simple approximation heuristics are developed, one of which guarantees solutions to be at most two times the optimum. For large‐scale problems, a simulated annealing heuristic that uses random neighborhood generation is provided. Computational experiments are conducted to test the algorithms. © 2006 Wiley Periodicals, Inc. Naval Research Logistics, 2007.     

Rendez‐vous search on a rectangular lattice

In a rendez‐vous search two or more teams called seekers try to minimize the time needed to find each other. In this paper, we consider s seekers in a rectangular lattice of locations where each knows the configuration of the lattice, the distribution of the seekers at time 0, and its own location, but not the location of any other. We measure time discretely, in turns. A meeting takes place when the two seekers reach the same point or adjacent points. The main result is that for any dimension of lattice, any initial distribution of seekers there are optimal strategies for the seekers that converge (in a way we shall make clear) to a center. © 2007 Wiley Periodicals, Inc. Naval Research Logistics, 2007     

互联网中文搜索引擎存在的问题与发展趋势探析

搜索引擎是互联网信息检索技术的核心,中文搜索引擎与国外相比存在着搜索引擎覆盖率低、查准率不高、检索精度差、更新速度慢、无法控制网络信息的动态变化、对信息内容难于控制和管理等问题。对此,提出了中文搜索引擎从单一查询工具向综合化、全功能服务发展;分类主题一体化,检索语言向自然语言发展;查询智能化,向多国化和多语种化发展;搜索引擎集成化、网络资源著录标准化及利用智能代理技术的信息检索等对策。     

求解布尔不可满足子式的消解悖论算法

求解布尔不可满足子式在超大规模集成电路设计与验证领域都具有非常重要的理论与应用价值,帮助EDA工具迅速定位错误与不一致。针对求解不可满足子式的非完全方法,提出了消解悖论与悖论解析树的概念,在此基础上提出一种启发式局部搜索算法。该算法根据公式的消解规则,采用局部搜索过程直接构造证明不可满足性的悖论解析树,而后递归搜索得到不可满足子式;算法中融合了布尔推理技术、动态剪枝方法及蕴含消除方法以提高搜索效率。基于随机测试集进行了实验对比,结果表明提出的算法优于同类算法。     

基于分布搜索的传感器指示交接协商机制

针对传感器指示交接技术中"指示任务怎样执行"的问题,根据指示交接的实际过程,研究了多传感器针对多个协作任务怎样进行更好调度的问题,提出了基于分布搜索的多维协商机制。通过协商函数的迭代更新,不断寻找综合效用最大的传感器调度策略,优化了指示交接过程中多传感器多任务调度过程。实例仿真结果表明,所提方法合理、有效,提高了传感器协作的效率,为指示交接的深入研究奠定了基础。