THE CSI JOURNAL ON COMPUTER SCIENCE AND ENGINEERING
Year:2008 | Volume:5 | Issue:2-4 (B)|Papers Count:7

In this paper, we have used a fuzzy net trained by the particle swarm optimization (PSO) technique for classification task of data mining.Fuzzy net uses a single layer neural network to reduce the complexities associated with multi-layer perceptron (MLP). In addition, we trained the network by using PSO to obtain optimal weight vectors, which reduces the problem of being trapped into a local minimum. From the simulation study, it is observed that the fuzzy swarm net (FSN) gives a promising result. Further, we compared the result obtained from FSN with multi-layer perceptron (MLP) radial basis function (RBF). The results obtained from FSN draws a clear edge between FSN and REF and a competitive result with MLP.

Reconfigurable computing offers advantages over traditional hardware and software implementation of computational algorithms. It is based on using reprogrammable devices, which can be reconfigured after fabrication to implement the desired algorithm.Reconfigurable computing systems can take advantage of hardware but still maintains the flexibility of software.Particular applications, including encryption, are specifically well suited to these systems. In this paper, we implemented multiple cryptographic algorithms, namely DES, LOKI, DESX, Biham-DES, and SnOES on a reconfigurable hardware so that each algorithm could be replaced by another with low reconfiguration overhead time. The hardware reconfiguration time is about 2.6 ms, which is comparable with the other systems overhead.The main reason for the lower reconfiguration time is that our architecture is partially reconfigurable. Our implementation resulted in a high flexibility but comparable ciphering rate in comparison with previous implementations on field programmable gate arrays (FPGAs). We achieved the ciphering rate of 14080 Mbps on Xilinx 2V6000-5 FPGA.

Process variation, despite introducing challenges in stability and power of SRAM memories, provides a new opportunity for leakage-power reduction: statistical simulation shows that the same SRAM cell leaks differently when storing 0 and 1; this difference is as high as 57% at 60mv variation of threshold voltage (Vth). Thus, leakage can be reduced if values with less leakage can be stored in the cells. We show applicability of this proposal by presenting a first technique for reducing instruction cache leakage: we reorder instructions within basic-blocks so as to match up the instructions with the less-leaky state of their corresponding cache cells.Experimental results show up to 12.51%, averaging 10.73%, leakage energy reduction at 60mv variation in Vth, and that this saving increases with technology scaling. Since intra-basic-block rescheduling does not affect instruction cache hit ratio, this reduction is provided with only a negligible penalty, in rare cases, in the data cache.

The surge of facilitating the ever-growing number of service classes in the contemporary Internet backbone has persuaded researches worldwide to search for suitable underlying network topologies with desirable properties. Even in the presence of speedy optical technologies, the arrangement of nodes in a network can be highly influential in boosting network performance.Among all the topologies investigated so far, k-ary n-cubes have been reported to be widely adopted in the literature. However in a hybrid network, where the network comprises of a mixture of different types of topologies, the type and size of k-ary n-cubes can greatly influence the performance factor of the network. The objective of this paper is twofold. Firstly, we introduce a set of new metrics that can be used to analyze the performance behavior of simple regular meshes with their wrap-around variants (2-D and 3-D torus) for the Optical Packet Switching (OPS) paradigm.We then propose a simple and yet efficient switch-level analytical model for service differentiation in OPS in the presence of wavelength converters (WCs). The proposed model is further extended to include the impact of Fiber Delay Lines (FDLs) as well. Simulations are conducted for both cases to corroborate the accuracy of the models presented.

One of the shortcomings of the existing clustering methods is their problems dealing with different shape and size clusters. On the other hand, most of these method sare designed for especial cluster types or have good performance dealing with particular size and shape of clusters. The main problem in this connection is how to define a dissimilarity criterion to make this algorithm capable of clustering general data, which include clusters of different shapes and sizes. Another important objective that must be considered is the computational complexity of any new algorithms.In this paper a new approach to fuzzy clustering is proposed in which a model for each cluster is estimated during learning.Gradually besides, dissimilarity metric for each cluster is defined, updated and used for the next step. In our approach, instead of associating a single cluster type to each cluster, we assume a set of possible cluster types for each cluster with different grades of possibility. Then, a truncation which can be expressed as an attention mechanism focuses on the most probable cluster types for each cluster. This selection step subsides the computational load dramatically while speeds up the clustering.The proposed clustering method which has the capability to deal with partial labeled data is implemented on two families of data, first in presence of partially labeled data, then with fully unlabeled data. Comparing the experimental results of this method with several important existing algorithms, demonstrates the superior performance of proposed method. The merit of this method is its ability to deal with clusters of different shape and size while it computes a fuzzy membership value to different shapes for each cluster.

In this paper we provide a process algebra to formally analyze ad hoc protocols above network layer. Mainly our focus in this paper is on formal modeling and analysis of routing protocols. We explain the semantics of our algebra using labeled transition systems and then define an equivalence relation between them. Finally we illustrate application of our algebra by a case study on cluster formation process of cluster based ad hoc routing protocol.

QoS support is necessary for the Internet new applications to meet end-to-end requirements such as bandwidth, delay, jitter and packet loss.Unfortunately, DiffServ architecture can only provide limited service assurance. Thus, a trade-off between simplicity of the implementation and strength of the service guarantees has to be made. In this paper, we have focused on DiffServ architecture to improve strength of service guarantees. We propose a new algorithm, using Adaptive Model Predictive Control as the controller to schedule differentiated buffers in routers. Simulation results show that the new algorithm can be used as a solution for the scheduling problem to guarantee QoS parameters like absolute delay, proportional delay and jitter.