Earticle

With the development of context aware computing, an adaptive, reflective, scalable, dynamic reconfigurable framework is urgently required. However, most current frameworks fail to meet new challenges in scalability and dynamic reconfiguration. These can be mainly attributed to two issues. First, a framework’s real hope for widespread adoption lies in its easy incorporation of existing third party resources. Second, the framework is required to better support the dynamic reconfiguration, which means framework behaviors should be altered according to the change of contexts. To address these issues, an Open Web Service based Multi-Agent framework (OWS-MA) is proposed and implemented. We illustrate the overall framework, elaborate on the individual goals of each part, and illuminate how each part responses the challenges. Our exploration discloses a promising way to the future developments of context aware application.

To have high performance scheduling mechanisms in grid computing, we need accurate methods for estimating parameters like jobs' wait time and run time. In this paper, we consider wait time prediction problem. Different regression techniques are examined on AuverGrid data set to predict wait time. To improve the quality of prediction, some extra features are proposed. Simulation results show that adding these features reduces prediction error between 13% and 60% in different methods. Results also show that K-nearest neighbor outperforms other regression techniques. We have compared the k-nearest neighbor method in both original and enriched data set with Last-M. K-Nearest neighbor in enriched data set outperforms both Last-M and K-nearest neighbor in original data set in accuracy and perfect prediction percentage.

Wireless Sensor Networks (WSNs) have been gaining growing interest in the past years, which has resulted in many proposals of their new applications. The main tasks of a sensor network include not only monitoring and measuring certain phenomena, but also delivering gathered data. Hence, every single node needs to be a wireless router aside from being only a simple sensing device. The idea and the attributes of a Wireless Sensor Network require designers to apply different techniques than those hitherto used in wired as well as in wireless networks. The dynamic progress in the field of WSNs entails the need of creating simulators that have more specific capabilities. Such simulation tools should allow researchers to verify new ideas and compare the proposed solutions in a virtual environment helping to avoid unnecessary, time-consuming or expensive hardware implementations. When beginning to work with Wireless Sensor Networks, it is important to choose a simulation environment that will be up to requirements and will allow the researcher to conduct experiments in a given area. Appropriately, the researcher faces a necessity of finding and getting oneself familiar with many simulators, often designed for radically different applications. This article presents a classification and a wide review the main features of various simulation platforms.

We introduce a new category of fuzzy neural networks with multiple-output based on fuzzy clustering algorithm, especially, fuzzy c-means clustering algorithm (FCM-based FNNm) for pattern classification in this paper. The premise part of the rules of the proposed networks is realized with the aid of the scatter partition of input space generated by FCM clustering algorithm. The partitioned local spaces describe the fuzzy rules and the number of the partitioned local spaces is equal to the number of clusters. Due to these characteristics, we may alleviate the problem of the curse of dimensionality. The consequence part of the rules is represented by polynomial functions with multiple-output for pattern classification. And the coefficients of the polynomial functions are learned by back propagation algorithm. To optimize the parameters of the proposed FCM-based FNNm we consider real-coded genetic algorithms. The proposed networks are evaluated with the use of numerical experimentation.

Grid technology allows resource sharing among several entities, but selecting the best resource to run a specific job remains one of its main problems. We try to solve this problem by predicting the performance of the resources for a given job and submit our job to best suited one. So an appropriate resource selection mechanism is crucial for the success of job execution and performance of Grid. The QoS parameter is key factor in our selection process. The QoS ensure that high quality requested jobs which may be a communicational or computational based must be submitted to high quality resources. In this paper we proposed a resource selection mechanism which worked in two phases, first phase filter the resources based upon their trust value and second phase predict the performance of all the host for a given job by considering QoS and rank the resources according to user criteria and submit the job to top most resource.

Information content plays an important role in measuring semantic similarity of concepts. The conventional way of IC obtained is through statistical analysis of corpora. Recently corpora–independent model has attracted great concern in this area. This paper analyzes the state-of-art IC models, highlights important related issues, and presents a novel IC model based on concepts’ topology in WordNet. Different from previous work, for a given concept, the depth itself, the number of its hyponyms, and the depth of every hyponym have been taken into considered. Experiment demonstrates that our approach is able to provide more accurate similarity evaluation and achieves significant performance than related works.

Cloud computing can provide facilities such as infrastructure and applications to clients on demand. Successful servicing of cloud paradigm necessitates accurate resource allocation. Whereas there are different workload types with different characteristics that should be supported by cloud computing, there is no any single solution that can allocate resources to all imaginable demands optimally. Therefore, there is a need to design specific solutions to allocate resources for each workload type. Based on that, in this paper, we focus on bag of tasks workload type and propose an idea to facilitate dynamic resource allocation for it. Technically, the proposed approach exploits users’ service level agreement parameters and classifies them. It controls utilization of servers to response users in a reasonable time. To validate the proposed approach, we evaluate it using the Monte Carlo simulation. Then, the experimental results are compared with two reference models, namely First Fit and Proportional Share. The proposed approach outperforms the reference models in terms of the total cost of resource allocation and response time of clients.

Since the volume of data generated by a scientific data experiment has grown exponentially, new scientific methods to analyze and organize the data are required. Hence, these methods need to be used effective infrastructure composed of computing resources that are used for pre-processing and post-processing data. The demanding requirement has led to development of methods to reduce the size of dataset and to apply a new programming model and its implementation like MapReduce. In this paper, we describe an empirical study for handling the dataset of a scientific data experiment to support data transformation, which is an essential phase to handling large-scale data in scientific data experiments. In this experiment we show a way to optimize the dataset written in netCDF by a data reduction as a sub-setting method and to process the dataset about tornado outbreak in the US by Hadoop, a MapReduce framework. These methods can be applied to pre-processing and post-processing in scientific data experiments.

Grid computing allows, controlled and coordinated resource sharing and problem solving in dynamic, multi- institutional virtual organizations. Grids are categorized in to two, as Computational grid and Data grid. In computing grid, allocating resources to the tasks based on its requirements and task scheduling is an important issue. Generally, the task scheduler assigns tasks to a suitable resource. The nodes with better performance would be assigned first for execution. When the load on a grid was high, and all resource with a better performance is assigned, other tasks have to be assigned to the resource with unsuitable performance. To improve the performance and throughput of the system, it is required to assign the tasks to a suitable and lower fault rate processor. Therefore, if a task is allocated to a resource without considering the performance factor, resource fault rate, the overall execution time will increases and throughput will decreases obviously. To solve this problem, a task scheduling algorithm that finds suitable and low fault rate resource for task execution is introduced in this paper. Experiments made on GridSim, shows that the proposed algorithm works well in dynamic environments.

Sink mobility has been proposed to optimize energy efficiency and lifetime for wireless sensor networks (WSNs). Many novel data retrieving algorithms or protocols are proposed based on sink mobility; they can be divided into two classes, namely path-unlimited and path-limited algorithms. In terms of application range, the path-limited algorithms are more applicable in many scenarios. In path-limited WSNs, the challenge is optimizing network energy utilization and lifetime based on the limited conditions. In this paper, we propose a novel data retrieving mechanism named Multiple Enhanced Specified-deployed Subsinks (MESS) mechanism for WSNs with path-limited mobile sink. We argue that deploying some enhanced wireless nodes, namely subsinks to retrieve data locally, and then mobile sink passes though these subsinks and harvests data from them. Simulation results show that MES mechanism can dramatically improve network lifetime compare to traditional mechanism.

The performance of decentralized detection in power constrained wireless sensor network is analyzed. By using the distributed processing schemes with the analog relay amplifier local processing, this system is proposed to be subjected to a total power constrains. And, the distributed nodes are assumed to perform analog relay amplifier local processing. Under such conditions the effect of the sensor to fusion communication by making use of the DS-CDMA concept. It evolves the fusion performance by using non-orthogonal sensor-to-fusion center communication, as opposed to the orthogonal communication. Also, to quantify the performance a circulant matrix theory is used to derive closed form asymptotic expressions. The presented results show the effect of each parameter on the total system performance.