基于记忆原理的人工鱼智能认知模型

将生物学的记忆原理引入到计算机动画中,提出了一种基于记忆原理的人工鱼智能认知模型。在该模型中,利用短时记忆元和长时记忆衰减和更新的特点及其与综合记忆的联系,建立了相应的数学模型,综合记忆值的大小决定了人工鱼对该事物的记忆强度,特征记忆库的更新清理功能有效的节省了系统的开销。实验结果证明,该模型在一定程度上提高了人工鱼的认知能力,具有很好的应用价值。     

基于竞争演化算法的TSP求解

针对NP完全问题的TSP问题，该文提出了一种属于启发式算法的竞争演化算法。并用构造能量函数的方法证明，用这种算法能使能量函数减小。最终必找到一条有意义的路径。微机仿真结果说明，这种算法能在较少的迭代步骤内找出一条较短的路径。     

一种面向堆操作程序内存安全性的域敏感内存模型

堆操作程序具有通过共享易变数据结构动态操纵堆内存单元的特性,使得内存安全性难以保证。针对这个问题,提出了一种域敏感的k-limit内存抽象模型,以支持动态调整抽象的粒度,取得静态分析在精度和效率上的平衡。分别从框架、性质、操作方面介绍了该内存模型,然后结合内存安全性的定义,在基于该模型的操作语义框架内定义了4种与内存安全性相关的错误类型,最后设计了基于该模型内存安全性检测的数据流迭代算法。     

A numerical method for determining monotonicity and convergence rate in iterative learning control

In iterative learning control (ILC), a lifted system representation is often used for design and analysis to determine the convergence rate of the learning algorithm. Computation of the convergence rate in the lifted setting requires construction of large N×N matrices, where N is the number of data points in an iteration. The convergence rate computation is O(N²) and is typically limited to short iteration lengths because of computational memory constraints. As an alternative approach, the implicitly restarted Arnoldi/Lanczos method (IRLM) can be used to calculate the ILC convergence rate with calculations of O(N). In this article, we show that the convergence rate calculation using IRLM can be performed using dynamic simulations rather than matrices, thereby eliminating the need for large matrix construction. In addition to faster computation, IRLM enables the calculation of the ILC convergence rate for long iteration lengths. To illustrate generality, this method is presented for multi-input multi-output, linear time-varying discrete-time systems.     

Reducing Off-Chip Memory Access via Stream-Conscious Tiling on Multimedia Applications

The iteration space of a loop nest is the set of all loop iterations bounded by the loop limits. Tiling the iteration space can effectively exploit the available parallelism, which is essential to multiprocessor compiling and pipelined architecture design. Another improvement brought by tiling is the better data locality that can dramatically reduce memory access and, consequently, the relevant memory access energy consumptions. However, previous studies on tiling were based on the data dependence, thus arrays without dependencies such as input arrays (data streams) were not considered. In this paper, we extend the tiling exploration to also accommodate those dependence-free arrays, and propose a stream-conscious tiling scheme for off-chip memory access optimization. We show that input arrays are as important, if not more, as the arrays with data dependencies when the focus is on memory access optimization instead of parallelism extraction. Our approach is verified on TI’s low power C55X DSP with popular multimedia applications, exhibiting off-chip memory access reduction by 67% on average over the traditional iteration space tiling.     

固态存储器短周期存取速度动态控制

由于传统控制方法出现数据误码、存储内存不足问题,导致控制性能变差,为了解决该问题,需对固态存储器短周期存取速度进行动态控制。根据短周期存取速度控制原理,采用均衡加重技术,设计具有针对性的数据传输介质来解决数据误码问题。利用NANDFLASH双平面交替编程的分时加载操作方式,扩大存储内存。使用二级缓存方式对扩大后的内存数据进行缓存读取,根据多线组合指令动态控制缓存读取结果。通过实验结果得出,该方法最低误码率可达到5%,控制性能较好。     

面向控制的分数阶微分模型的快速数值计算

求解分数阶控制系统的关键在于如何快速精确地计算分数阶微分.针对短记忆法和变步长记忆法的计算精度和计算复杂性顾此失彼的矛盾,本文提出了一种恒权重记忆法,它不舍弃历史数据,而是采用常值权重后全部记忆.在每个后继的采样周期,只需把新的数据简单叠加到历史数据上来考虑,从而极大地提高了计算精度和降低了计算复杂性,且有效地化解了两...     

具有全局记忆的LF蚁群聚类算法

针对传统的LF蚁群聚类算法中存在的收敛速度慢,蚂蚁空载导致的资源浪费以及易陷入局部最优等问题,提出了一种蚁群改进算法。算法初期采用直接分配原则,直接将蚂蚁随机放在数据对象上,并生成随机的全局记忆,在聚类时负载蚂蚁移动受到全局记忆的指导,利用余弦相似度判断最相似的记忆中心,并向该记忆中心移动,全局记忆在一次迭代完成后更新。当蚂蚁拾起数据对象失败时,为了减少蚂蚁再一次的随机移动所带来的资源浪费,采用相异原则将蚂蚁移动到下一个数据对象上。改进的算法在UCI数据集Iris、Wine、Glass和Robotnavigation上进行验证,算法在保证原有算法准确率的基础上明显提高了收敛速度。     

伪卫星时钟同步算法的研究

在伪卫星自主组网定位的时钟同步问题中,卡尔曼滤波(Kalman Filter,KF)算法因迭代时间短而被广泛应用,但是其精度不高.相反地,粒子滤波(Particle Filter,PF)算法精度高,但是其迭代时间长.在此基础上,提出一种混合优化算法(Hybrid Optimizing Algorithm,HOA),该算...     

嵌入式实时操作系统中实现MBUF

MBUF是TCP／IP中的概念，本文从它在嵌入式实时系统中的应用，讨论了它的基本原理和在嵌入式实时操作系统中的实现方法以及相关的内存管理支持。     

An improved SCF iteration scheme

An easily implementable iteration scheme guaranteeing convergence under very lenient conditions is developed and implemented into the MNDO-program package. Calculations done using large atomic clusters obtained from modelling diamond and silicon surfaces show that this scheme is, with respect to the requirements of computer time and memory, superior to some of the other iteration methods successfully used in quantum chemical calculations.     

Parallel single grid and multigrid solution of industrial compressible flow problems

The Euler and Navier-Stokes equations with a k-ε turbulence model are solved numerically in parallel on a distributed memory machine IBM SP2, a shared memory machine SGI Power Challenge, and a cluster of SGI workstations. The grid is partitioned into blocks and the steady state solution is computed using single grid and multigrid iteration. The multigrid algorithm is analyzed leading to an estimate of the elapsed time per iteration. Based on this analysis, a heuristic algorithm is devised for distributing and splitting the blocks for a good static load balance. Speed-up results are presented for a wing, a complete aircraft and an air inlet.     

Efficient solution of fluid flow using the generalised conjugate grandient algorithm on a transputer-based machine

The discretisation of the equations governing fluid flow gives rise to coupled, quasi-linear and non-symmetric systems. The solution is usually obtained by iteration using a guess-and-correct procedure where each iteration aims to improve the solution of the previous step. Each step or outer iteration of the process involves the solution of nominally linear algebraic systems. These systems are normally solved using methods based on the Gauss-Seidel iteration—such as the TDMA. However, these methods generally converge very slowly and can be very time consuming for realistic applications. In this paper, these equations are solved using the Generalised Conjugate Gradient (GCG) algorithm with a simple-to-implement Gauss-Seidel-based preconditioner on a distributed memory message-passing machine. We take advantage of the fact that only tentative improvements to the flow-field are sought during each iteration and study the convergence behaviour of the parallel implementation on a multi-processor environment.     

Hermite算子在图像配准中的应用

基于Euler-Lagrange方程的医学图像弹性配准算法在临床中有着广泛的应用,它是医学图像处理的研究热点问题。传统算法采用差分近似计算图像的梯度误差较大,本文利用Hermite微分算子计算图像的梯度信息;并在迭代过程中对外力进行平滑约束。通过对模拟和真实数据的实验表明改进方法不仅具有较高的计算精度,而且具有很强的鲁棒性,适合于医学图像的弹性配准。     

基于DM642的光纤中心定位算法优化

大规模的数据存取是制约数字信号处理器系统效率的主要因素之一。在传像光纤束成像系统中,为提高出端光纤束在光纤中心定位过程中数据存取的效率,提出DM642平台下的光纤中心定位优化方法。在存储器配置方面,设置二级储存器缓存与片内随机存储器的模式,预取部分片外数据到片内随机存储器。在数据传输方面,设计数据存储动态偏移量,自适应迭代,避免重复读入。实验结果表明,优化后的算法实现了光纤中心的定位,时间开销降低了1/4,提高了系统的运行效率。     

融合长短时记忆机制的机械臂多场景快速运动规划

针对快速扩展随机树(rapid-exploration random trees,RRT)算法难以有效解决多场景环境下的机械臂快速运动规划问题,提出一种融合长短时记忆机制的快速运动规划算法.首先,采用高斯混合模型(Gaussian mixture models,GMM)在规划的初始阶段通过随机采样构建环境的场景模型,并利用该模型进行碰撞检测,以提高运动规划效率;然后,根据人类的记忆机制原理,对多场景的不同GMM按照即时记忆、短期记忆和长期记忆进行存储,并通过场景匹配算法实现不同场景GMM的快速自适应提取,提高对变化环境的适应能力;最后,通过在Matlab以及ROS仿真环境下6自由度柔性机械臂的运动规划仿真实验对所提出的算法进行验证.实验结果表明,所提出算法可以快速提取场景的记忆信息,有效提高多场景环境下的运动规划效率,具有较强的适应性.     

Dedicated microprocessors for realtime identification of multivariable systems

In this paper, an algorithm is proposed for identifying multivariable systems in state-space form from noisy data, which is suitable for implementation on dedicated microprocessor systems. The proposed algorithm uses the normalized stochastic approximation criterion which reduces the computational complexity and memory requirements. It is shown that the overall performance of the proposed stochastic approximation algorithm when using a dedicated microprocessor with fixed point arithmetic is superior to the extended least-squares method in terms of memory requirements, execution speed per iteration, and the estimation results.     

基于预拷贝的虚拟机动态内存迁移机制改进

针对内存预拷贝过程中迁移时间较长和内存页反复重传的特点,改进传统的内存动态迁移机制,引入马尔科夫预测模型,提出基于脏页概率预测的工作集测定算法.利用脏页的历史操作访问情况预测其下一轮迭代被修改的概率,只传输预测概率较低的页.实验结果表明,该算法缩短了迁移总时间和停机时间,能有效支持虚拟机动态迁移.     

基于分解转移矩阵的PageRank迭代计算方法

本文提出了一种基于分解转移矩阵的PageRank的迭代计算方法。该方法对PageRank理论模型进一步推导,把其Markov状态转移矩阵进行了分解,从而降低存储开销和计算复杂度,减少I/O需求,使得PageRank计算的工程化实现更为简单。实验表明1 700多万的网页2.8亿条链接,可以在30秒内完成一次迭代,内存需求峰值585MB,可以满足工程化应用的需求。