P. Spirakis
University of Liverpool, Computer Science, Faculty Member
Research Interests:
Research Interests:
Research Interests:
Stackelberg games [15] may model the interplay among an authority and rational individuals that selfishly demand resources on a large-scale network. In such a game, the authority (Leader) of the network is modeled by a distinguished... more
Stackelberg games [15] may model the interplay among an authority and rational individuals that selfishly demand resources on a large-scale network. In such a game, the authority (Leader) of the network is modeled by a distinguished player. The selfish users (Followers) are modeled by the remaining players. It is well known that selfish behavior may yield a Nash Equilibrium with cost arbitrarily higher than the optimum one, yielding unbounded Coordination Ratio or Price of Anarchy (PoA) [7, 13]. Leader plays his strategy first assigning a portion of the total demand to some resources of the network. Followers observe and react selfishly assigning their demand to the most appealing resources. Leader aims to drive the system to an a posteriori Nash equilibrium with cost close to the overall optimum one [4, 6, 8, 10]. Leader may also be eager for his own rather than system’s performance [2, 3]. A Stackelberg game can be seen as a special, and easy [6] to implement, case of Mechanism Design. It avoids the complexities of either computing taxes or assigning prices, or even designing the network at hand [9]. However, a central authority capable to control the overall demand on the resources of a network may be unrealistic in networks which evolute and operate under the effect of many and diversing economic entities. A realistic way [4] to act centrally even in large nets could be via Virtual Private Networks (VPNs) [1]. Another flexible way is to combine such strategies with Tolls [5, 14]. A dictator controlling the entire demand optimally on the resources surely yields PoA = 1. On the other hand, rational users do prefer a liberal world to live. Thus, it is important to compute the optimal Leader strategy which controls the minimum of the resources (Price of Optimum) and yields PoA D 1. What is the complexity of computing the Price of Optimum? This is not trivial to answer, since the Price of Optimum depends crucially on computing an optimal Leader strategy. In particular, [6] proved that computing the optimal Leader strategy is hard.
Research Interests:
Research Interests:
Research Interests:
We present here a high performance parallel free-text retrieval m ethod for multiple text queries using the vector space model. Our method employs the fat-tree area universal routing network for connecting the processors of the parallel m... more
We present here a high performance parallel free-text retrieval m ethod for multiple text queries using the vector space model. Our method employs the fat-tree area universal routing network for connecting the processors of the parallel m achine. We provide a theoretical analysis of our technique which shows it t o b e superior to parallel t ext retrieval based
Research Interests:
Research Interests:
Research Interests:
Research Interests:
... 5. H. Chan, A. Perrig, and D. Song, “Random Key Predistribution Schemes for Sensor Networks”, Proceedings of the IEEE Symposium of Privacy and Security”, 11–14 May, pp. 197–213, 2003. 6. A. CF. Chan and ES Rogers ...
Research Interests: Computer Science, Graph Theory, Modeling, Localization, Computer Security, and 15 moreTheoretical Computer Science, Distributed System, Mobile Computing, Mobile Ad Hoc Network, Wireless Network, Key Management, Sensor Network, Ad hoc network, Graph, Secure Communication, Sensor Array, Boolean Satisfiability, Wireless Sensor Network, Wireless Ad Hoc Network, and Graph Model
Research Interests:
In a number of situations, it is necessary for two agents who may have never communicated in the past to, jointly, create a shared information item which can serve as a basis for subsequent protocols that the agents may wish to execute... more
In a number of situations, it is necessary for two agents who may have never communicated in the past to, jointly, create a shared information item which can serve as a basis for subsequent protocols that the agents may wish to execute (e.g., negotiation or encryption protocols). One way to create this shared piece of information is to have the two agents start with one random bit string each and then engage in a protocol that enables them to transform, gradually, bit differences (in their strings) into bit agreements. In a previous work, an efficient protocol was proposed which was based on the use of the Extended Golay error-correcting code in order to locate and “correct” bit differences. In this work we generalize this protocol in order to use any generic error-correcting code and derive theoretical performance bounds on the efficiency, based on the characteristics of the employed code. The proposed generalized protocol is fair, in that the final strings (which have the same bits in the majority of positions) depend on the strings possessed by both agents while each agent contributes to the same degree in the formation of these strings. Finally, the proposed protocol is lightweight (both computationally and with respect to message exchanges) and, thus, can be implemented in embedded systems and resource limited devices.
Research Interests:
ABSTRACT In this paper, we propose simple protocols for enabling two communicating agents that may have never met before to extract common knowledge out of any initial knowledge that each of them possesses. The initial knowledge from... more
ABSTRACT In this paper, we propose simple protocols for enabling two communicating agents that may have never met before to extract common knowledge out of any initial knowledge that each of them possesses. The initial knowledge from which the agents start, may even be independent of each other, implying that the two agents need not have had previous access to common information sources. In addition, the common knowledge extracted upon the termination of the protocols depends, in a fair way, on the (possibly independent) information items initially known, separately, by the two agents. It is fair in the sense that there is a negotiation between the two agents instead of one agent forcing the other to conform to its own knowledge. These protocols, may be extended in order to support security applications where the establishment of a common knowledge is required. Moreover, the implementation of the protocols leads to reasonably small code that can also fit within resource limited devices involved in any communication network while, at the same time, it is efficient as simulation results demonstrate.
Research Interests:
ABSTRACT
Research Interests:
A classical problem in scheduling theory is to find an optimal nonpreem-tive sched~e for a collection of unit length tasks subject to precedence con-straints. We are given n tasks to be executed on m processors. Each task requires exactly... more
A classical problem in scheduling theory is to find an optimal nonpreem-tive sched~e for a collection of unit length tasks subject to precedence con-straints. We are given n tasks to be executed on m processors. Each task requires exactly one unit of execution and can run on any ...
Research Interests:
Research Interests:
The wide growth of Internet users has led the developers of popular Web sites to adopt the solution of substituting the Web servers by clusters of Web servers, in order to correspond efficiently to the high rate of requests received. They... more
The wide growth of Internet users has led the developers of popular Web sites to adopt the solution of substituting the Web servers by clusters of Web servers, in order to correspond efficiently to the high rate of requests received. They also applied load balancing techniques to distribute the workload among them. In this paper we present a general design of load balancing strategies in a cluster of Web servers, which uses a proxy per server to collect the incoming requests. To accomplish the best possible workload distribution, we use an update process, which can either be centralized or distributed. We apply two sender-initiated load balancing algorithms, based on the distinction between the centralized and the distributed update process, and analyze their performance results. We conclude that the distributed case behaves better under high loaded conditions.
Research Interests:
This paper shows that the problem of testing whether a graph $G$ contains an induced subgraph of vertex (edge) connectivity at least $k$ is P-complete for any fixed $k \geqslant 3$. Moreover, if $k_{\max } $ is the largest vertex (edge)... more
This paper shows that the problem of testing whether a graph $G$ contains an induced subgraph of vertex (edge) connectivity at least $k$ is P-complete for any fixed $k \geqslant 3$. Moreover, if $k_{\max } $ is the largest vertex (edge) connectivity of any subgraph of $G$, ...
Research Interests:
Research Interests:
Research Interests: Computer Science, Parallel Computing, Technology, Evaluation, Queueing theory, and 14 morePerformance Evaluation, Mathematical Sciences, Asynchronous Communication, Performance Model, ATM networks, Model Evaluation, Steady state, asynchronous Transfer Mode, Analytical Model, Interconnection, Multiprocessing, Exact solution methods, Multistage Interconnection Networks, and Multistage Interconnection Network (MIN)
Research Interests:
ABSTRACT Open and Distance Learning (ODL) can be regarded as learning by means of Telematics (i.e. the combination of means of telecommunications, information technology and multimedia), by which way, among other things:
Research Interests:
Research Interests:
... overhead. [2] present the AODV (Ad Hoc On De-mand Distance Vector routing) protocol that also uses a demand-driven route establishment procedure. ... inconceiv-Page 3. Analysis and Experimental Evaluation 101 able ...
Research Interests: Computer Science, Distributed Computing, Mobile Ad Hoc Network, Computer Network, Random Walk, and 12 moreTheoretical Analysis, Message Passing, Routing Protocol, Experimental Evaluation, Mobile Network, Experimental Validation, Experimental Measurement, Wireless Communication, Wireless Ad Hoc Network, Mobile Host, Mobile Station, and Algorithm Engineering
Research Interests:
Research Interests: Computer Science, Algorithms, Distributed Computing, Randomization, Scheduling, and 13 moreResource Allocation, Distributed System, Probability, DISTRIBUTION, Communication System, Data Bases, Resource Management, Real Time, Efficient Resource Management in Cloud Computing, Distribution, Distributed Algorithm, Probability Distribution Functions, and Parallel Communication
Research Interests:
Research Interests:
Research Interests:
Research Interests:
Research Interests:
Research Interests:
In this work, we study the stability of the FIFO (First-In-First-Out) protocol in the context of Adversarial Queueing Theory. As an important intermediate step, we consider dynamic capacities, where each network link capacity may... more
In this work, we study the stability of the FIFO (First-In-First-Out) protocol in the context of Adversarial Queueing Theory. As an important intermediate step, we consider dynamic capacities, where each network link capacity may arbitrarily take on values in the two-valued set of integers f1; Cg for C > 1 being the high capacity (a parameter). In this context: (1) We construct a FIFO network of only eight nodes which is already unstable at rate r = 0:41. This is the current record for instability of FIFO over networks of flxed-size (independent of r). (2) For every r > 0 we then construct a FIFO network (whose size increases with 1 r ) which is unstable at rate r. Subsequently, we show how to simulate the particular FIFO network in (2) above with dynamic capacities 1, C, in order to produce a FIFO network with all link capacities being equal, while preserving instability thresholds. Hence, we eventually show our main result: FIFO can become unstable in the usual model of unit...
Research Interests: Mathematics, Computer Science, Routing, Queueing theory, Stability, and 4 moreFIFO, ECCC, Queue, and First In First Out
Research Interests:
Research Interests:
Research Interests:
Research Interests:
Research Interests: Mathematics, Computer Science, Combinatorics, Wireless Sensor Networks, Construction, and 14 moreTheoretical Computer Science, Computing, Probability, Network, Variance, Mathematical Sciences, Algorithmics, Efficient Algorithm for ECG Coding, Intersection, Randomized Algorithm, Wireless Sensor Network, Regular graph, Probability Distribution, and Independent set
A brief survey on the most recent results concerning the simulation of PRAMs on DMMs is presented, along with the design of a Shared Memory Simulator (SMS) that is based on a distributed environment provided by many realistic... more
A brief survey on the most recent results concerning the simulation of PRAMs on DMMs is presented, along with the design of a Shared Memory Simulator (SMS) that is based on a distributed environment provided by many realistic architectures, based on the DMM model. The Shared Memory Simulation Frame (SMS-Frame) will give the opportunity of executing parallel algorithms designed for a variant of PRAM, over a purely distributed working environment, that will be c onsidered to comply with the DMM model. The reason why the models of PRAM and DMM have been chosen, is that t he former has proved to be the most popular cost model i n the parallel algorithms community, while the latter is an abstracted model based on message passing, that is very close to, or is provided through specialized interfaces by the majority of the vendors of parallel machines.