George Cybenko

Dynamic load balancing for distributed memory multiprocessors

In this paper we study diffusion schemes for dynamic load balancing on message passing multiprocessor networks. One of the main results concerns conditions under which these dynamic schemes converge and their rates of convergence for arbitrary topologies. These results use the eigenstructure of the iteration matrices that arise in dynamic load balancing. We completely analyze the hypercube network by explicitly computing the eigenstructure of its node adjacency matrix. Using a realistic model of interprocessor communications, we show that a diffusion approach to load balancing on a hypercube multiprocessor is inferior to another approach which we call the dimension exchange method. For a d-dimensional hypercube, we compute the rate of convergence to a uniform work distribution and show that after d + 1 iterations of a diffusion type approach, we can guarantee that the work distribution is approximately within e-* of the uniform distribution independent of the hypercube dimension d. Both static and dynamic random models of work distribution are studied. o

Tracking a moving object with a binary sensor network

In this paper we examine the role of very simple and noisy sensors for the tracking problem. We propose a binary sensor model, where each sensors value is converted reliably to one bit of information only: whether the object is moving toward the sensor or away from the sensor. We show that a network of binary sensors has geometric properties that can be used to develop a solution for tracking with binary sensors and present resulting algorithms and simulation experiments. We develop a particle filtering style algorithm for target tracking using such minimalist sensors. We present an analysis of fundamental tracking limitation under this sensor model, and show how this limitation can be overcome through the use of a single bit of proximity information at each sensor node. Our extensive simulations show low error that decreases with sensor density.

AGENT TCL: targeting the needs of mobile computers

Mobile computers have become increasingly popular as users discover the benefits of having their electronic work available at all times. Using Internet resources from a mobile platform, however, is a major challenge. Mobile computers do not have a permanent network connection and are often disconnected for long periods. When the computer is connected, the connection is often prone to sudden failure, such as when a physical obstruction blocks the signal from a cellular modem. In addition, the network connection often performs poorly and can vary dramatically from one session to the next, since the computer might use different transmission channels at different locations. Finally, depending on the transmission channel, the computer might be assigned a different network address each time it reconnects. Mobile agents are one way to handle these unforgiving network conditions. A mobile agent is an autonomous program that can move from machine to machine in a heterogeneous network under its own control. It can suspend its execution at any point, transport itself to a new machine, and resume execution on the new machine from the point at which it left off. Agent Tcl is a mobile agent system whose agents can be written in Tcl, Java, and Scheme. Agent Tcl has extensive navigation and communication services, security mechanisms, and debugging and tracking tools. We focus on Agent Tcls architecture and security mechanisms, its RPC system, and its docking system, which lets an agent move transparently among mobile computers, regardless of when they are connected to the network.

The Numerical Stability of the Levinson-Durbin Algorithm for Toeplitz Systems of Equations

The numerical stability of the Levinson-Durbin algorithm for solving the Yule-Walker equations with a positive-definite symmetric Toeplitz matrix is studied. Arguments based on the analytic results of an error analysis for fixed-point and floating-point arithmetics show that the algorithm is stable and in fact comparable to the Cholesky algorithm. Conflicting evidence on the accuracy performance of the algorithm is explained by demonstrating that the underlying Toeplitz matrix is typically ill-conditioned in most applications.

Mobile Agents for Distributed Information Retrieval

A mobile agent is an executing program that can migrate during execution from machine to machine in a heterogeneous network. On each machine, the agent interacts with stationary service agents and other resources to accomplish its task. Mobile agents are particularly attractive in distributed information-retrieval applications. By moving to the location of an information resource, the agent can search the resource locally, eliminating the transfer of intermediate results across the network and reducing end-to-end latency. In this chapter, we first discuss the strengths of mobile agents, and argue that although none of these strengths are unique to mobile agents, no competing technique shares all of them. Next, after surveying several representative mobile-agent systems, we examine one specific information-retrieval application, searching distributed collections of technical reports, and consider how mobile agents can be used to implement this application efficiently and easily. Then we spend the bulk of the chapter describing two planning services that allow mobile agents to deal with dynamic network environments and information resources: (1) planning algorithms that let an agent choose the best migration path through the network, given its current task and the current network conditions, and (2) planning algorithms that tell an agent how to observe a changing set of documents in a way that detects changes as soon as possible while minimizing overhead. Finally, we consider the types of errors that can occur when information from multiple sources is merged and filtered, and argue that the structure of a mobile-agent application determines the extent to which these errors affect the final result.

Computing thr minimum eigenvalue of a symmetric positive definite Toeplitz matrix

A method for computing the smallest eigenvalue of a symmetric positive definite Toeplitz matrix is given. It relies solely upon the Levinson–Durbin algorithm. The procedure involves a combination of bisection and Newtons method. Good starting values are also shown to be obtainable from the Levinson–Durbin algorithm.

Reducing quantum computations to elementary unitary operations

Quantum computations are intimately connected with unitary operators. This article shows that standard techniques from numerical linear algebra can be used to represent quantum computations as sequences of simple quantum operations, called quantum Givens operators, on single quantum bits.

Mobile agents: the next generation in distributed computing

Mobile agents are programs that can move through a network under their own control, migrating from host to host and interacting with other agents and resources on each. We argue that these mobile, autonomous agents have the potential to provide a convenient, efficient and robust programming paradigm for distributed applications, particularly when partially connected computers are involved. Partially connected computers include mobile computers such as laptops and personal digital assistants as well as modem connected home computers, all of which are often disconnected from the network. We describe the design and implementation of our mobile agent system, Agent Tcl, and the specific features that support mobile computers and disconnected operation. These features include network sensing tools and a docking system that allows an agent to transparently move between mobile computers, regardless of when the computers connect to the network.

Cybersecurity Strategies: The QuERIES Methodology

QuERIES offers a novel multidisciplinary approach to quantifying risk associated with security technologies resulting in investment-efficient cybersecurity strategies. R esearchers can use the QuERIES methodology to rigorously determine, for the first time, appropriate investment levels and strategies for the protection of intellectual property in complex systems. As a result, it can have a significant and immediate impact on the protection of critical IP, including weapons systems and chip designs, complex computer software, and databases containing personal and financial information. In this paper, initial testing of QuERIES in small-scale, realistic scenarios, were performed with results that suggest the methodology can significantly improve risk assessments in complex systems under attack by rational and capable adversaries. Such systems include software, hardware, and data critical to national security and industrial competitiveness. Consequently, it is believed that QuERIES has wide applicability within both the DoD and private sectors.

The Traveling Agent Problem

Abstract. This paper considers a sequencing problem which arises naturally in the scheduling of software agents. We are given n sites at which a certain task might be successfully performed. The probability of success is pi at the ith site and these probabilities are independent. Visiting site i and trying the task there requires time (or some other cost metric) ti whether successful or not. Latencies between sites i and j are lij, that is, the travel time between those two sites. Should the task be successfully completed at a site then any remaining sites do not need to be visited. The Traveling Agent Problem is to find the sequence which minimizes the expected time to complete the task. The general formulation of this problem is NP-Complete. However, if the latencies are constant we show that the problem can be solved in polynomial time by sorting the ratios ti/pi according to increasing value and visiting the sites in that order. This result then leads to an efficient algorithm when groups of sites form subnets in which latencies within a subnet are constant but can vary across subnets. We also study the case when there are deadlines for solving the problem in which case the goal is to maximize probability of success subject to satisfying the deadlines. Applications to mobile and intelligent agents are described.