树形全光网络中波长转换器配置算法

配置有限数量的波长转换器使网络阻塞率最低,是全光网络中需要解决的一个关键问题.通过考虑网络的直径、中心以及节点和链路的通信量,采用网络分解和迭代的方法,提出树形网络中基于赋权直径的波长转换器配置算法、基于节点加权中心的波长转换器配置算法,以及基于光路加权中心的波长转换器配置算法.算法演示表明,提出的3个算法总是将波长转换器放置在阻塞率较高的节点上,从而大大降低网络整体阻塞率.     

高层次综合的实有化逻辑图自动生成关键技术的研究

研究的逻辑图自动生成系统是“九五”微电子重点科技预研项目的成果“实用化专用集成电路高层次自动化设计系统－Talent系统”的子系统,该系统通过对电路网表有效的识别与划分、逻辑单元的布局及互连信号线的布线等,自动生态具有一定逻辑功能且布局美观规范的逻辑原理图。该文重点研究逻辑图自动生成实用化过程中的关键技术,将人工智能基于规则的知识表示与形式化算法相结合,提出有效的自动布局方法;应用模式识别理论和方法解决逻辑图的自动布线问题,提出基于决策树的通道分配方法,并设计一套完整的规则体系;进而,文中给出详细的划分模型,并在此基础上,结合逻辑原理图的特点实现了两种有效的划分算法：（1）种子生成的构造式划分算法;（2）迭代改进划分算法。其中算法（1）的设计思想主要源于贪婪构造,而算法（2）则对经典的最小分割划分算法Kernigham－Lin算法进行改进。基于上述研究实现的逻辑图自动生成系统,能够在很短的时间内生成美观规范、可读性好的逻辑图,为整个系统实用化做出了贡献。     

云计算中能耗和性能感知的虚拟机优化部署算法

优化虚拟机部署是数据中心降低能耗的一个重要方法。目前大多数虚拟机部署算法都明显地降低了能耗,但过度虚拟机整合和迁移引起了系统性能较大的退化。针对该问题,首先构建虚拟机优化部署模型。然后提出一种二阶段迭代启发式算法来求解该模型,第一阶段是基于首次适应下降装箱算法,提出一种虚拟机优化部署算法,目标是最小化主机数;第二阶段是提出了一种虚拟机在线迁移选择算法,目标是最小化待迁移虚拟机数。实验结果表明,该算法能够有效地降低能耗,具有较低的服务等级协定(SLA)违背率和较好的时间性能。     

岛式FPGA线长驱动快速布局算法

传统的FPGA布局箅法需要花费大量时间,影响了FPGA物理设计效率.为了在保证布局质量的前提下缩短布局时间,提出一种岛式FPGA快速布局算法.首先考虑终端传输的迭代二划分,然后进行最小费用流初始布局和低温模拟退火的布局优化.在每一个划分层次中,考虑了线网的终端对线网权重的影响;对于每一个划分的区域,使用最小费用流来确定初始的布局;在布局的最后阶段使用低温模拟退火来提高初始布局的质量.实验结果表明,该算法布局结果的质量高、速度快.     

动态虚拟盘阵数据分布算法

尽管当今的磁盘等外存储设备容量增加得很快,但还是无法满足用户应用程序的需要;在性能上,外存储设备已经成为计算机系统的瓶颈;为此,在集群环境下,将分布式的外设构成动态虚拟盘阵系统是一种较好的解决方案,而数据分布算法是动态虚拟研究的一项重要内容。也就是说,采用优化的数据分布算法,使得盘阵的性能和容量随盘阵的扩展而扩展。研究的主要工作是综述以往对动态盘阵数据分布算法,并对以往SCADDAR算法进行了扩充,提出了D／H(Double／Halve)数据分布算法。     

Cuckoo search based resource optimization of datacenters

With advancements in virtualization technology, datacenters are often faced with the challenge of managing large numbers of virtual machine (VM) requests. Due to this large amount of VM requests, it has become practically impossible to search all possible VM placements in order to find a solution that best optimizes certain design objectives. As a result, managers of datacenters have resorted to the employment of heuristic optimization algorithms for VM placement. In this paper, we employ the cuckoo search optimization (CSO) algorithm to solve the VM placement problem of datacenters. Firstly, we use the CSO to optimize the datacenter for the minimization of the number of physical machines used for placement. Secondly, we implement a multiobjective CSO algorithm to simultaneously optimize the power consumption and resource wastage of the datacenter. Simulation results show that both CSO algorithms outperform the reordered grouping genetic algorithm (RGGA), the grouping genetic algorithm (GGA), improved least-loaded (ILL) and improved FFD (IFFD) methods of VM placement.     

基于遗传算法的布局问题求解

布局问题是VLSI布图设计中的重要问题,传统的方法很难得到满意的解答。针对该问题,文中提出了一个基于遗传算法的求解方法,并将它应用到VLSI门阵列模式布局设计中,构造了一个较完备的综合布局目标函数,引入并定义了通道拥挤度。实验结果表明,所提出的方法能在较短的时间内提供优化解,为VLSI布局设计提供了一种新的思路。     

Placement Strategies for Internet-Scale Data Stream Systems

Optimally assigning streaming tasks to network machines is a key factor that influences a large data-stream-processing system's performance. Although researchers have prototyped and investigated various algorithms for task placement in data stream management systems, taxonomies and surveys of such algorithms are currently unavailable. To tackle this knowledge gap, the authors identify a set of core placement design characteristics and use them compare eight placement algorithms. They also present a heuristic decision tree that can help designers judge how suitable a given placement solutions might be to specific problems.     

An augmented Lagrangian method for VLSI global placement

By ignoring some cell overlaps, global placement computes the best position for each cell to minimize the wirelength. It is an important stage in very large scale integration (VLSI) physical design, since circuit performance heavily depends on the placement results. In this paper, we propose an augmented Lagrangian method to solve the VLSI global placement problem. In the proposed method, a cautious dynamic density weight strategy is used to balance the wirelength objective and the density constraints, and an adaptive step size is used to obtain a trade-off between runtime and solution quality. The proposed method is tested on the IBM mixed-size benchmarks and the International Symposium on Physical Design 2006 placement contest benchmarks. Experimental results show that our global placement method outperforms the state-of-the-art placement approaches in terms of solution quality on most of the benchmarks.     

A performance guarantee heuristic for electronic components placement problems including thermal effects

In this work, Benders decomposition algorithm is used to deal with a computer motherboard design problem. Amongst all the possible formulations for the component placement problem, the chosen one creates an instance of the extensively studied Quadratic Assignment Problem (QAP). This problem arises as a great challenge for engineers and computer scientists. The QAP inherent combinatorial structure makes the most efficient optimization algorithms to exhibit low performance for real size instances. It is also considered here the important addition of linear costs. This approach is directly responsible for the performance gain presented by our decomposition method. Coupled with the placement problem, it is under investigation the maximum temperature rising on the board surface. In order to solve the Energy Conduction Equation the Finite Volume Method is implemented, becoming possible to derive a secondary quality solution criterion. A set of test instances is then solved and the corresponding results are reported.     

Evolutionary Algorithms for Minimax Problems in Robust Design

Many robust design problems can be described by minimax optimization problems. Classical techniques for solving these problems have typically been limited to a discrete form of the problem. More recently, evolutionary algorithms, particularly coevolutionary optimization techniques, have been applied to minimax problems. A new method of solving minimax optimization problems using evolutionary algorithms is proposed. The performance of this algorithm is shown to compare favorably with the existing methods on test problems. The performance of the algorithm is demonstrated on a robust pole placement problem and a ship engineering plant design problem.     

基于视觉强化学习的数字芯片全局布局方法

在数字芯片后端设计中,全局布局需要同时兼顾线长与合法化,是一个组合优化问题。传统的退火算法或者遗传算法耗时且容易陷入局部最优,目前强化学习的解决方案也很少利用布局的整体视觉信息。为此,提出一种融合视觉信息的强化学习方法实现端到端的全局布局。在全局布局中,将电路网表信息映射为多个图像级特征,采用卷积神经网络(convolutional neural network, CNN)和图卷积网络(graph convolutional network, GCN)将图像特征和网表信息相融合,设计了一整套策略网络和价值网络,实现对全局布局的全面分析和优化。在ISPD2005基准电路上进行实验,结果证明设计的网络收敛速度加快7倍左右,布局线长减少10%～32%,重叠率为0%,可为数字芯片全局布局任务提供高效合理的方案。     

DrPlace:基于深度学习的可布线性驱动布局算法

在集成电路物理设计的布局阶段,针对基于深度学习的布局算法结果可布线性较差的问题,在开源的DREAMPlace算法的基础上提出并实现了一种基于深度学习的可布线性驱动布局算法DrPlace.算法模型在总体上设计并实现了布局器的整体框架,集成了基于深度学习的可布线性驱动总体布局、可布线性驱动的合法化和详细布局.总体布局过程中...     

Resource placement in Cartesian product of networks

The utilization of the limited resources of a multiprocessor or multicomputer system is a primary performance issue which is crucial for the design of many scheduling algorithms. While many of the existing parallel machines benefit from a regular product network topology, almost none of the previous resource placement techniques have come to recognize and exploit this inherent regularity. This paper introduces several novel algorithms for deriving resource placement schemes in product networks based on the assumption of perfect resource placement in their underling basic graphs. Our techniques use known schemes for the basic networks as their building blocks for deploying the resource placement scheme in the product network. This seriously cuts down the expenses required for deploying and rescaling the network. In particular, we propose some efficient algorithms for adjacency placement in a product of k$k$ heterogeneous graphs. Furthermore, we extend our approach and present algorithms for distant resource placement in product networks.     

Two optimization methods for complicated structures with multiparameter elements based on generalized condensation

A high-order condensation expression is proposed to generalize Duffin's condensation formula. Based on the higher-order condensation expression, two primal algorithms of the second-order are constructed to solve GGP (generalized geometric programming) problems. To overcome the difficulty of solving the structural optimization with multiparameter elements, two methods are proposed and implemented in a computer program. The first method takes one geometrical property of the cross-section as one component of the design variables for each element and adopts SQP (sequential quadratic programming) to solve the optimization model. The second method takes geometrical sizes of the cross-section as the design variables for each element and adopts primal algorithms of GGP problems to solve the optimization model.     

大规模网络存储系统的数据布局算法研究

可扩展和可靠的数据管理是实现大规模网络存储系统的关键技术,需要一个能够自适应存储规模变化、公平、冗余、高可用的数据布局算法。本文根据不同的设计目标以及存储环境的同构和异构特性,分析了四类布局算法:同构的布局算法、非冗余的异构布局算法、冗余的异构布局算法和高可用的布局算法。对现有布局算法存在的问题进行了讨论,并指出了未来的研究方向。     

Thermal-aware floorplanner for 3D IC,including TSVs,liquid microchannels and thermal domains optimization

3D stacked technology has emerged as an effective mechanism to overcome physical limits and communication delays found in 2D integration. However, 3D technology also presents several drawbacks that prevent its smooth application. Two of the major concerns are heat reduction and power density distribution. In our work, we propose a novel 3D thermal-aware floorplanner that includes: (1) an effective thermal-aware process with three different evolutionary algorithms that aim to solve the soft computing problem of optimizing the placement of functional units and through silicon vias, as well as the smooth inclusion of active cooling systems and new design strategies, (2) an approximated thermal model inside the optimization loop, (3) an optimizer for active cooling (liquid channels), and (4) a novel technique based on air channel placement designed to isolate thermal domains have been also proposed. The experimental work is conducted for a realistic many-core single-chip architecture based on the Niagara design. Results show promising improvements of the thermal and reliability metrics, and also show optimal scaling capabilities to target future-trend many-core systems.     

分层并行遗传算法和遗传复合形算法及其应用

基于复合形算法、遗传算法、分层和并行思想,设计了一种求解复杂多目标、多约束和多变量工程优化问题的分层并行遗传或复合形算法,编制了界面友好和计算可靠性高的VC++软件。对于一类复杂三多工程综合优化问题,进行了遗传算法、复合形算法、分层并行遗传算法和分层并行遗传复合形算法的大量计算,结果表明：分层并行遗传算法计算效率最高;为解决复杂的三多工程综合优化问题提供了有效的可行方法。     

Discrete cooperative particle swarm optimization for FPGA placement

Particle swarm optimization (PSO) is a stochastic optimization technique that has been inspired by the movement of birds. On the other hand, the placement problem in field programmable gate arrays (FPGAs) is crucial to achieve the best performance. Simulated annealing algorithms have been widely used to solve the FPGA placement problem. In this paper, a discrete PSO (DPSO) version is applied to the FPGA placement problem to find the optimum logic blocks and IO pins locations in order to minimize the total wire-length. Moreover, a co-operative version of the DPSO (DCPSO) is also proposed for the FPGA placement problem. The problem is entirely solved in the discrete search space and the proposed implementation is applied to several well-known FPGA benchmarks with different dimensionalities. The results are compared to those obtained by the academic versatile place and route (VPR) placement tool, which is based on simulated annealing. Results show that both the DPSO and DCPSO outperform the VPR tool for small and medium-sized problems, with DCPSO having a slight edge over the DPSO technique. For higher-dimensionality problems, the algorithms proposed provide very close results to those achieved by VPR.