基于CUDA的AES并行算法优化

为提升高级加密标准(AES)的加密性能,利用显卡的通用计算能力,在统一计算设备架构(CUDA)平台上实现AES的128位、192位和256位3个版本的GPU并行算法,并提出优化的AES并行算法。在考虑块内线程数量、共享存储器容量和总块数的基础上,根据分块最优值的经验数据指导AES算法在GPU上的最优分块。实验结果表明,与未优化的AES并行算法相比,该算法的3个版本在Nvidia Geforce G210显卡上的加密速度分别提高5.28%,14.55%和12.53%,而在Nvidia Geforce GTX460显卡上的加密速度分别提高12.48%,15.40%和15.84%,且能更好地对SSL数据进行加密。     

基于改进离散粒子群禁忌算法的并行测试任务调度

针对并行测试任务调度问题特点,提出了一种解决并行测试任务调度的离散粒子群禁忌算法,设计了符合并行测试任务调度特点的粒子初始化编码方式、粒子位置转移公式以及禁忌搜索邻域。通过与禁忌搜索相结合,避免了离散粒子群算法早熟问题。通过实验以及算法对比,表明提出的算法是快速有效的,能够很好地解决并行测试任务调度问题,求得测试总时间最短的任务调度序列。     

自适应混合多目标分布估计进化算法

针对多目标分布估计算法全局收敛性较弱的缺陷,提出了一种自适应混合多目标分布估计进化算法。其基本思想是:在多目标分布估计算法中引入全局收敛性较强的差分进化算法,当函数变化率较大时,用分布估计算法产生新种群;当函数变化率较小即算法可能陷入局部收敛时,用差分进化算法产生新种群。理论分析和数值实验结果表明,这种混合算法不仅具有良好的全局收敛性,而且解的分布性和均匀性较没有考虑目标函数变化率的混合多目标分布估计算法也有了一定程度的提高。     

原材料价格波动下钢铁生产库存多期优化模型

考虑钢铁企业钢材生产受钢材需求和生产变动成本波动的双重影响,在预测基础上,建立钢材生产库存多期动态优化模型。由于模型涉及多种钢材及多个时段,属于大规模问题,求解困难,且为了避免粒子群算法陷入局部最优,提出采用模拟退火与粒子群组合智能算法对模型进行求解。最后通过钢铁企业L的案例,结果表明算法具有较强的收敛性和适用性,模型可用于解决钢铁企业多期生产实际问题。     

凸轮磨恒线速加工过程曲线优化

为了提高凸轮轴的加工效率和加工精度,开发可用于凸轮轴磨削修整系统的动态优化模块。根据凸轮轴恒线速磨削原理,在Matlab中完成了数学模型的建模,采用三次样条曲线拟合插值法对加工过程曲线进行了优化,从而得到了波动更小的速度曲线并减少了恒线速度磨削所引起的联动轴响应问题。实验表明,该方法很好地实现了在二次磨削中对凸轮轴表面轮廓的修整,提高了凸轮轴的表面质量。     

基于属性约简的MPI运行时参数优化

目前MPI实现提供上百种可供调整的运行时参数,但默认的运行时参数不能发挥程序应有的性能。为了给应用程序提供接近最优的运行时参数,本文提出基于粗糙集理论中的属性约简建立优化模型。首先通过属性约简方法为不同基准程序产生参数组合值,再根据未知应用程序的程序特征通过所建立的优化模型为其预测参数值。实验表明,该方法可以有效地为未知应用程序预测参数值,并且比使用默认参数程序性能提升约20%。     

实数编码量子共生演算法及其在云任务调度中的应用

针对共生演算法收敛慢和易陷入局部最优的问题,结合量子遗传算法理论,提出一种实数编码的量子共生演算法(real-coded quantum symbiotic organisms search,RQSOS)。首先依据三角模糊数提出差异度概念,并依此构造一个以自变量向量的分量和一对概率幅为等位基因的三倍染色体,使一条染色体携带更多信息并增强解的多样性;然后提出一种基于阿基米德螺旋线的探索学习模式,加强对解空间的探索精度;最后使用共生演算法更新差异度值并依据差异度值对种群进行学习和变异操作,促使整个种群快速向最优方向进化且减小了陷入局部最优的概率。利用数值优化问题和云任务调度问题对算法进行验证,仿真结果表明,RQSOS算法在收敛速度和寻优能力上均有明显提升,是一种可行有效的算法。     

Improving cache locality for GPU-based volume rendering

We present a cache-aware method for accelerating texture-based volume rendering on a graphics processing unit (GPU). Because a GPU has hierarchical architecture in terms of processing and memory units, cache optimization is important to maximize performance for memory-intensive applications. Our method localizes texture memory reference according to the location of the viewpoint and dynamically selects the width and height of thread blocks (TBs) so that each warp, which is a series of 32 threads processed simultaneously, can minimize memory access strides. We also incorporate transposed indexing of threads to perform TB-level cache optimization for specific viewpoints. Furthermore, we maximize TB size to exploit spatial locality with fewer resident TBs. For viewpoints with relatively large strides, we synchronize threads of the same TB at regular intervals to realize synchronous ray propagation. Experimental results indicate that our cache-aware method doubles the worst rendering performance compared to those provided by the CUDA and OpenCL software development kits.     

化工搅拌器的参数优选

本文提出了确定化工用搅拌器搅拌时间的一种简单试验方法,并用正交表安排寻找最佳搅拌参数的试验和整理试验数据,得到了有实用价值的结果。     

MORM: A Multi-objective Optimized Replication Management strategy for cloud storage cluster

Effective data management is an important issue for a large-scale distributed environment such as data cloud. This can be achieved by using file replication, which efficiently reduces file service time and access latency, increases file availability and improves system load balancing. However, replication entails various costs such as storage and energy consumption for holding replicas. This article proposes a multi-objective offline optimization approach for replica management, in which we view the various factors influencing replication decisions such as mean file unavailability, mean service time, load variance, energy consumption and mean access latency as five objectives. It makes decisions of replication factor and replication layout with an improved artificial immune algorithm that evolves a set of solution candidates through clone, mutation and selection processes. The proposed algorithm named Multi-objective Optimized Replication Management (MORM) seeks the near optimal solutions by balancing the trade-offs among the five optimization objectives. The article reports a series of experiments that show the effectiveness of the MORM. Experimental results conclusively demonstrate that our MORM is energy effective and outperforms default replication management of HDFS (Hadoop Distributed File System) and MOE (Multi-objective Evolutionary) algorithm in terms of performance and load balancing for large-scale cloud storage cluster.