Dynamics of knowledge creation in global participatory science communities: open innovation communities from a network perspective

Significant attention has been given to advancing cyber-infrastructures to support virtual engineering and science communities based on the proposition that virtual organizations can more effectively create and leverage knowledge due to diverse information, skills, and resources to enhance capacity to innovate. Yet, relatively little is known about desirable organizing processes in virtual open science communities. To this end, a simulation-based exploratory study is conducted to better understand the conditions that confer increased rates of innovation in such socio-technical systems. Three types of open science communities are identified and simulated using agent simulation as a method of inquiry. Simulation results show that centrality, as a measure of degree of connectedness, correlates with innovation output in exploratory and service communities up to a point. Also, utility-oriented communities have social network structures with low density and high centrality, which suggest high potential for innovation.     

Optimal Server Acquisition in Open Queueing Networks

Several production and flexible manufacturing systems can naturally be modelled using queueing networks. In this paper, we consider the problem of acquiring servers for the nodes of an open queueing network, so as to optimize the steady-state mean virtual system parameters subject to a budget constraint. A partial enumeration scheme and a heuristic method have been proposed to solve this problem. Empirical results based on randomly generated test problems are used to identify a class of problems for which the heuristic performs well.     

Off-line and on-line guaranteed start-up delay for Media-on-Demand with stream merging

We address the problem of designing efficient algorithms for media-on-demand in systems that use stream merging. In the stream merging model the receiving bandwidth of clients is larger than the playback bandwidth, and clients can buffer parts of the transmission to be played back later. Our goal is to minimize the required server bandwidth for a given guaranteed start-up delay and uninterrupted playback. We construct an efficient O(n) optimal off-line algorithm for a time horizon that is composed of n time slots, where the length of one slot is the guaranteed start-up delay. Our algorithm works for either clients with receiving bandwidth twice as much as the playback bandwidth or for receiving bandwidth equal to the server bandwidth, independent of the clients buffer size. We describe an on-line delay guaranteed algorithm that operates without knowledge of the time horizon size, and show that it performs asymptotically close to the optimal off-line algorithm. The on-line algorithm is simpler to implement than previously proposed on-line stream merging algorithms, and empirically performs well when the intensity of client arrivals is high.     

Where Forward-Looking and Backward-Looking Models Meet

The present paper begins by deriving an instantaneous formulation for the backward-looking (reinforcement based learning) satisfaction balance model of Gray and Tallman (1984). This model is then used to generate interactional data from four simulated agents in a network interaction experiment. Because this initial model does not generate stable interaction structures in the network experiment, it is altered step by step in the direction of a forward-looking (agent with goals) model that has been shown to generate such stable interaction structures. The purpose of the modifications are to learn what aspects of the forward-looking model are needed to evolve a stable interaction structure, and to learn how these aspects may be incorporated into a model that remains essentially reinforcement based.     

A traffic shaping model for optimizing network operations

The management of technology in multi-service computer networks, such as university networks, has become a challenge with the explosive growth of entertainment oriented peer-to-peer (P2P) traffic. Traffic shaping is one of the tools used to manage bandwidth to improve system performance by allocating bandwidth between P2P and non-peer-to-peer (NP2P) traffic. We present a model for traffic shaping and bandwidth management that considers the trade-offs from allocating different amounts of bandwidths for different application categories and use data from a university network. The current policy allocates varying bandwidths over the day to P2P and NP2P traffic to reflect the importance of not letting entertainment based traffic choke the network during the day time at the expense of the more important traffic, such as Web traffic. We highlight the difficulties in obtaining data in the form required for analysis, and the need to estimate demand for allocations not covered by current policy. We present a goal programming model for this estimation task. We also model the traffic shaping problem as a Markov decision process and develop an algorithm for determining the optimal bandwidth allocation to maximize the utility of all users. Finally we use a numerical example to illustrate our approach.     

云计算环境资源监控系统研究

为保障云计算平台的可靠性,基于Ganglia,采用Eucalyptus开源云平台,提出了云计算平台资源监控系统架构.从总体结构、监控信息处理流程等多个方面对监控系统进行了详细阐述,提出了动态更换虚拟机簇头的算法,并对算法与原型系统进行了验证和性能评测.分析与评测结果表明,所提系统能够适用于云计算环境资源信息的实时监测与预警,具备系统开销较低,用户接口友好的特点,能准确反映系统、虚拟机的负载信息,有助于云计算系统服务可靠性的提升.簇头动态更换算法能及时发现失效的簇头并更换为更健壮的虚拟机,保证了监测系统的正常运行.     

Reactive model for autonomous vehicles formation following a mobile reference

The emergence of collective motion in nature is ubiquitous and can be observed from colonies of bacteria to flocks of birds. The scientific community is interested in understanding how the local interactions drive the crowd toward global behaviors. This paper presents an agent-based reactive model for groups of vehicles that aims to make the formation to follow a moving reference, represented as a virtual agent. The model is called reactive because the agents do not keep previous information but only respond to the current system state. Moreover, they only communicate with their close neighbors, limited by their sensory radius, except with the virtual agent that can be seen by everyone at the whole time. The aim of the model is to group the agents around the virtual agent while it moves to desirable directions. We solve the inverse problem of parameter estimation in order to drive the model toward specific objectives. This task is performed with the Generalized Extremal Optimization (GEO) algorithm, and the results are tested with path planning scenarios.     

The combinatorial bandwidth packing problem

In managing a telecommunications network, decisions need to be made concerning the admission of requests submitted by customers to use the network bandwidth. The classical bandwidth packing problem requires that each request submitted by a customer use network resources to establish a one-to-one connection involving one single pair of nodes. We extend the problem to the more practical case where each request submitted by a customer to use the network resources includes a set or combination of calls. This extension suggests that each request requires one-to-many or many-to-many connections to be established between many communicating node pairs. The extension has applications in many important areas such video conferencing and collaborative computing. The combinatorial nature of the requests makes the admission decision more complex because of bandwidth capacity limitations and call routing difficulties. We develop an integer programming formulation of the problem and propose a procedure that can produce verifiably good feasible solutions to the problem. The results of extensive computational experiments over a wide range of problem structures indicate that the procedure provides verifiably good feasible solutions to the problem within reasonable computational times.     

VIRTUAL SYSTEM METHOD FOR SYSTEM TREE

ＶＩＲＴＵＡＬＳＹＳＴＥＭＭＥＴＨＯＤＦＯＲＳＹＳＴＥＭＴＲＥＥ￥ＺＨＥＮＧＺＨＯＮＧＧＵＯ（ＤｅｐａｒｔｍｅｎｔｏｆＰｒｏｂａｂｉｌｉｔｙａｎｄＳｔａｔｉｓｔｉｃｓ，ＢｅｉｊｉｎｇＵｎｉｖｅｒｓｉｔｙ）Ｂｅ１Ｊｉｎｇ１００８７１，Ｃｈｉｎａ．）（Ｐ...     

加权支持向量机在证券指数预测中的研究

支持向量机是基于统计学习理论的新一代学习机器.它使用结构风险最小化原则,运用核技巧,较好地解决了学习问题.本文提出了一种基于支持向量机的加权算法,并将其应用于证券,指数预测.与径向基神经网络相比较,加权支持向量机表现出了良好的性能.     

Level‐based heuristics and hill climbing for the antibandwidth maximization problem

The antibandwidth maximization problem aims to maximize the minimum distance of entries of a sparse symmetric matrix from the diagonal and as such may be regarded as the dual of the well‐known bandwidth minimization problem. In this paper, we consider the feasibility of adapting heuristic algorithms for the bandwidth minimization problem to the antibandwidth maximization problem. In particular, using an inexpensive level‐based heuristic, we obtain an initial ordering that we refine using a hill‐climbing algorithm. This approach performs well on matrices coming from a range of practical problems with an underlying mesh. Comparisons with existing approaches show that, on this class of problems, our algorithm can be competitive with recently reported results in terms of quality while being significantly faster and applicable to much larger problems. Copyright © 2012 John Wiley & Sons, Ltd.     

基于LSTM-GA混合模型的患者预约排队策略优化

在医疗运作管理领域,合理的资源分配能够帮助更多的患者尽早就医,降低患者病情恶化和死亡的风险。本文设计了预约排队策略对患者占有资源的顺序进行分配,建立了基于长短时记忆(Long Short Term-Memory, LSTM)神经网络和遗传算法(Genetic Algorithm, GA)的混合模型以优化排队策略。首先利用大数据和深度学习分析患者到达和医院服务情况,建立LSTM神经网络学习数据特征并预测未来数据,相比于排队论常用的随机分布方法取得了更好的效果.其次设计了基于排队系统仿真的排队策略优化算法,利用改进GA得到最优排队策略。实证研究表明,文本的方法可以明显降低患者的等待时间,最高可达59%。最后对排队策略进行敏感性分析,结果表明排队策略有效作用于仿真的各个时段。     

Robust connection acceptance control for ATM networks with incomplete source information

Consider an ATM resource where the statistics of each bursty traffic source are not precisely known. How should the resource decide when to accept fresh connections? We address this problem for the case where each source behaves as a simple “On-Off” model and only the peak bandwidth is known. A connection acceptance control algorithm is proposed that relies on the choice of a boundary. The performance of this algorithm is analysed using an approximation based on separation of timescales. The slowly changing call process behaves like the queue process at an infinite server queue, where the arrival process to the queue is dictated by the fast moving burst process. A technique for finding a suitable boundary is also given, based on the Bayesian principle. The algorithm, together with the proposed boundary, could be used in a network to control connection acceptance at each resource along a virtual path.     

Modeling and optimizing a vendor managed replenishment system using machine learning and genetic algorithms

Using a supply chain network, we demonstrate the feasibility, viability, and robustness of applying machine learning and genetic algorithms to respectively model, understand, and optimize such data intensive environments. Deployment of these algorithms, which learn from and optimize data, can obviate the need to perform more complex, expensive, and time consuming design of experiments (DOE), which usually disrupt system operations. We apply and compare the behavior and performance of the proposed machine learning algorithms to that obtained via DOE in a simulated Vendor Managed Replenishment system, developed for an actual firm. The results show that the models resulting from the proposed algorithms had strong explanatory and predictive power, comparable to that of DOE. The optimal system settings and profit were also similar to that obtained from DOE. The virtues of using machine learning and evolutionary algorithms to model and optimize data rich environments thus seem promising because they are automatic, involving little human intervention and expertise. We believe and are exploring how they can be made adaptive to improve parameter estimates with increasing data, as well as seamlessly detecting system (and therefore model) changes, thus being capable of recursively updating and reoptimizing a modified or new model.     

Call admission control schemes and ATM network topological design

Call admission control criteria are not only important for call admission control itself, but also can be an important input to network topological design. In this paper, we show the difference in terms of network cost incurred by adopting different call admission control schemes in network topological design. We compare two call admission control schemes. Scheme 1 uses equivalent bandwidth as its call admission control criterion and Scheme 2 is based on modeling the volatility of call traffic using Reflected Brownian Motion. Though Scheme 2 increases the complexity of network topological design, it can give lower network costs. Our experimental results show that for the same traffic mix, the network cost can be as little as 10% and as much as 35% lower when Scheme 2 is used instead of Scheme 1. The differences between the pair of resulting networks suggests that network topological design can be used as one of the criteria for choosing the call admission control scheme.     

云环境下面向复杂资源需求的虚拟机能效部署研究

针对云环境下在线虚拟机部署这一矢量装箱问题进行了研究,提出了多维空间划分模型和在线虚拟机能效部署算法OEEVMP。多维空间划分模型可以引导虚拟机部署,避免多维资源的不均衡利用;基于此模型,提出的OEEVMP算法在物理机运行数量局部最优和全局最优之间取得均衡,从而提高虚拟机部署能效。通过仿真实验,将OEEVMP算法与MFFD算法进行了对比,实验结果验证了所提算法的可行性和有效性。最后,对控制模型的两个参数进行了分析,给出了最佳的参数组合。     

WAYFINDER: parallel virtual machine reallocation through A* search

Modern virtual machine (VM) management software enables consolidation of VMs for power savings or load-balancing for performance. While existing literature provides various methods for computing a better load-balanced, or consolidated goal state, it fails to adequately suggest the best path from the system’s current state to the desired goal allocation. This paper discusses an approach to efficient path finding in VM placement problems for cloud computing environments of moderate scale with results indicating the solution is reasonable for managing hundreds of VMs. We present an overview of known approaches to dynamic VM placement and discuss their shortcomings with respect to dynamic reallocation. We then describe a novel design and implementation of a heuristic search algorithm to determine optimal sequential migration plans to transition from a given VM-to-host allocation to an arbitrary desired allocation state. We then elaborate nuances of A* application to this domain along with our simulation-based validation approach. Finally, this work demonstrates a novel and highly effective technique for exploiting migration parallelism in order to rapidly achieving VM reallocation convergence suitable for continual workload rebalancing in cloud data centers.     

Matching performance objectives for open and closed workloads by consolidation and replication

One of the most crucial task during the design of a computing infrastructure is the decision about the proper amount of equipments required to handle a specific workload while satisfying a set of performance objectives. This problem is emphasized even more in actual computer infrastructure such as clouds, where the user can provision the resources very easily thanks to the use of virtual machines. If the system has to handle a low workload, resources can be consolidated together to reduce the costs. If however the workload is very high, resources must be replicated to gain an acceptable service level. In this paper we derive the impact on several performance indexes for both consolidation and replication when considering both open and closed workloads. In particular, we present an analytical model to determine the best consolidation or replication options that match given performance objectives specified through a set of constraints. Depending on the particular type of workload and constraints, we present either closed form expressions, heuristics or an iterative algorithm to compute the minimum number of resources required.     

A perspective on artificial intelligence: Learning to learn

The classical approach to the acquisition of knowledge in artificial intelligence has been to program the intelligence into the machine in the form of specific rules for the application of the knowledge: expert systems. Unfortunately, the amount of time and resources required to program an expert system with sufficient knowledge for non-trivial problem-solving is prohibitively large. An alternative approach is to allow the machine tolearn the rules based upon trial-and-error interaction with the environment, much as humans do. This will require endowing the machine with a sophisticated set of sensors for the perception of the external world, the ability to generate trial actions based upon this perceived information, and a dynamic evaluation policy to allow it to measure the effectiveness of its trial actions and modify its repertoire accordingly. The principles underlying this paradigm, known ascollective learning systems theory, have already been applied to sophisticated gaming problems, demonstrating robust learning and dynamic adaptivity.The fundamental building block of a collective learning system is thelearning cell, which may be embedded in a massively parallel, hierarchical data communications network. Such a network comprising 100 million learning cells will approach the intelligence capacity of the human cortex. In the not-too-distant future, it may be possible to build a race of robotic slaves to perform a wide variety of tasks in our culture. This goal, while irresistibly attractive, is most certainly fraught with severe social, political, moral, and economic difficulties.This paper was given as an invited talk on the 12th Symposium on Operations Research, University of Passau, September 1987.     

Load balancing optimization of telecommunication networks with two differentiated services

The Differentiated Services architecture is a scalable solution to provide differentiated Quality of Service. In this paper, we address the network load balancing optimization of such networks based on bandwidth differentiation between two services. We define the optimization problem as an Integer Programming model and propose a heuristic algorithm based on GRASP with Path Relinking. We present computational results showing that (i) good quality solutions can be computed and (ii) proper load balancing can efficiently obtain service differentiation.