A desired load distribution model for scheduling of unrelated parallel machines

Scheduling problems concern the allocation of limited resources over time among both parallel and sequential activities. Load balancing has been adopted as an optimization criterion for several scheduling problems. However, in many practical situations, a load-balanced solution may not be feasible or attainable. To deal with this limitation, this paper presents a generic mathematical model of load distribution for resource allocation, called desired load distribution (DLD). The objective is to develop a DLD model for scheduling of unrelated parallel machines that can be used both in centralized resource management settings and in agent-based distributed scheduling systems. The paper describes the proposed DLD model in details, presents a dynamic programming based optimization algorithm for the proposed model, and then discusses its application to agent-based distributed scheduling.     

Progressive filling partitioning and mapping algorithm for Spark based on allocation fitness degree

The job execution mechanism of Spark was analyzed,task efficiency model and Shuffle model were established,then allocation fitness degree (AFD) was defined and the optimization goal was put forward.On the basis of the model definition,the progressive filling partitioning and mapping algorithm (PFPM) was proposed.PFPM established the data distribution scheme adapting Reducers’ computing ability to decrease synchronous latency during Shuffle process and increase cluster the computing efficiency.The experiments demonstrate that PFPM could improve the rationality of workload distribution in Shuffle and optimize the execution efficiency of Spark.     

需求引领空间设计提升价值——以上海市宝山区体育中心设计为例

马斯洛需求层次理论把人类需求分成生理、安全、社交、尊重和自我实现五类，依次由较低层次到较高层次。据此理论，不同的空间形态及相互组合能够满足不同层次的心理需求，进而促进不同类型活动的发生及商业业态的发展，最终形成城市综合体育中心“以馆养馆”“以商养体”的局面，推动并保持项目的持续发展。     

Parallel Globally Consistent Normal Orientation of Raw Unorganized Point Clouds

A mandatory component for many point set algorithms is the availability of consistently oriented vertex‐normals (e.g. for surface reconstruction, feature detection, visualization). Previous orientation methods on meshes or raw point clouds do not consider a global context, are often based on unrealistic assumptions, or have extremely long computation times, making them unusable on real‐world data. We present a novel massively parallelized method to compute globally consistent oriented point normals for raw and unsorted point clouds. Built on the idea of graph‐based energy optimization, we create a complete kNN‐graph over the entire point cloud. A new weighted similarity criterion encodes the graph‐energy. To orient normals in a globally consistent way we perform a highly parallel greedy edge collapse, which merges similar parts of the graph and orients them consistently. We compare our method to current state‐of‐the‐art approaches and achieve speedups of up to two orders of magnitude. The achieved quality of normal orientation is on par or better than existing solutions, especially for real‐world noisy 3D scanned data.     

基于 MPI ＋OpenMP 的红外弱小目标检测并行计算

为有效监控红外弱小目标运动的全过程，必须采用多个波段同时探测，但是多波段探测必然带来计算时间的大幅增长，无法满足实际应用中对目标检测实时性的要求。针对这一问题，本文提出一种基于MPI＋OpenMP的层次化并行方法，充分利用消息传递模型和共享存储模型的优势，并基于多处理器节点集群进行测试。实验结果表明，该并行程序在保证相同的检测概率的情况下加速比达到8．61，极大地提高了目标检测的效率。     

Syntactic compression codes at the zero entropy point

Syntactic compression codes compress the tree, which is the syntax of a binary source message. The ones considered here originate from image processing. The syntactic trees usedhave a constant valency and their binary labels distinguish whether the source substring derived from a node is completely zero or not. We compress them simply by deleting some redundant subtrees.

These codes fall into a theoretically new class of codes which is wider than the classical ones. They are here studied in the neighborhood of a zero of the binary entropy function. There, their behavior is close to that of an infinite run length encoding and the optimum valency is three. Finally, we open a problem, related with automata theory, which perhaps could provide a further link between Information Theory and Algorithmic Information Theory.     

The parallel odd-even merge algorithm

This paper briefly describes the implementation of the odd-even merge algorithm on a parallel MIMD computer and discusses its computational complexity.     

低功率YAG激光-TIG电弧复合焊接镁合金薄板工艺

以2 mm厚AZ31B变形镁合金薄板为研究对象开展低功率脉冲YAG激光-TIG电弧复合热源焊工艺研究,分析了激光与电弧的能量匹配对焊缝成形的影响规律.结果表明,镁合金薄板低功率YAG激光-TIG电弧复合热源焊接过程中,激光能量与电弧能量之间的相互匹配将直接影响焊缝的表面成形,获得理想焊接接头的工艺参数区间相对较狭小.为使焊缝成形均匀连续,焊接过程中需要对焊缝背面采用氩气进行保护,当保护气体流量为5～10 L/min时获得了具有最佳性能的焊接接头,其拉伸载荷达到镁合金母材的95%以上.     

天津轧三2×1260 m~3高炉喷煤系统设计特点及运行实践

介绍了天津轧三钢铁有限公司2×1 260 m~3高炉喷煤系统的设计特点,讨论了煤粉计量、喷吹罐压力控制等问题。该喷煤系统采用三罐并列、喷吹主管加分配器的直接喷吹形式,系统喷吹能力大,单座高炉最大可达29 t/h;三罐并列形式实现了充压氮气的回收,延长了设备使用寿命。一年多的生产实践表明,该设计合理,布局紧凑,系统运行平稳,吨铁喷煤比控制在150 kg,有效地降低了生产成本。     

Analytic Function Expansion Nodal (AFEN) Method in Hexagonal-Z Three-Dimensional Geometry for Neutron Diffusion Calculation

The analytic function expansion nodal (AFEN) method in hexagonal-z geometry is described focusing on its unique features, including the use of node-interface flux moments. Multigroup extension based on matrix function theory and coarse group rebalance (CGR) acceleration are also described. The COREDAX code implementing the AFEN method is verified testing on the VVER-440 benchmark problem, a “simplified” VVER-1000 benchmark problem, and the SNR-300 benchmark problem.