Vol. 30 · No. 4

Distributed Real-Time Thread is an important concept for distributed real-time systems. Distributed Threads are schedulable entities with an end-to-end deadline that transpose nodes, carrying their scheduling context. In each node, the thread will be locally scheduled according to a local deadline, which is defined by a deadline partitioning algorithm. Mechanisms for predicting the missing of deadlines are fundamental if corrective actions are incorporated for improving system quality of service. In this work, a mechanism for predicting missing deadlines is proposed and evaluated through simulation. In order to illustrate the main characteristics of the proposed mechanism, experiments will be presented taking into account different scenarios of normal load and overload. Simulations show that the deadline missing prediction mechanism proposed presents good results for improving the overall performance and availability of distributed systems.

Refined descriptive sampling is designed to improve upon the descriptive sampling method for experimentation in simulation. The former reduces significantly the risk of sampling bias generated by descriptive sampling and eliminates its problem related to the sample size. In this paper, we propose an optimal parallel Monte Carlo simulation algorithm using refined descriptive sampling and evaluate in parallel architecture, performance measures of a stable M/M/1 queueing system, a Pert network and the Newsboy problem.

Fuzzy clustering is useful to mine complex and multi-dimensional data sets, where the members have partial or fuzzy relations. Among the various developed techniques, fuzzy-C-means (FCM) algorithm is the most popular one, where a piece of data has partial membership with each of the pre-defined cluster centers. Moreover, in FCM, the cluster centers are virtual, that is, they are chosen at random and thus might be out of the data set. The cluster centers and membership values of the data points with them are updated through some iterations. On the other hand, entropy-based fuzzy clustering (EFC) algorithm works based on a similarity-threshold value. Contrary to FCM, in EFC, the cluster centers are real, that is, they are chosen from the data points. In the present paper, the performances of these algorithms have been compared on four data sets, such as IRIS, WINES, OLITOS and psychosis (collected with the help of forty doctors), in terms of the quality of the clusters (that is, discrepancy factor, compactness, distinctness) obtained and their computational time. Moreover, the best set of clusters has been mapped into 2-D for visualization using a self-organizing map (SOM).

The vehicle routing problem (VRP) is famous as a nondeterministic polynomial-time hard problem. This study proposes a centroid-based heuristic algorithm to solve the capacitated VRP in polynomial time. The proposed algorithm consists of three phases: cluster construction, cluster adjustment, and route establishment. At the cluster construction phase, the farthest node (customer) among un-clustered nodes is selected as a seed to form a cluster. The notion of the geometrical centre of a cluster is introduced in this study to be utilized at the cluster construction and the cluster adjustment phases. The proposed algorithm has a polynomial time complexity of O(n2.2). Experimental results on Augerat benchmark dataset show that the proposed 3-phase approach can result in smaller distances than the Sweep algorithm in more cases.

A business process management system (BPMS) offers the facility to define new processes or update the existing processes. However, exceptional or non-routine tasks require the intervention of domain experts or generation of the situation specific resolution process. This paper assumes that sufficient amount of business process exception handling cases are stored in the process repository. Since the retrieval of the best exception handling process requires good understanding about the exceptional situation, context awareness is an important issue. To facilitate the representation of the exceptional situation and to enable the selection of the best exception handling process, we adopted the `situation variable' and `decision variable' construct. A case example for exception handling in the e-commerce delivery process is provided to illustrate how the proposed construct works. We applied the C5.0 algorithm to build the optimum search tree.

This paper describes the model and software tools for virtualization of the Grid execution environment. The basic assumption of this study is that Virtual Grid (VG) is constructed dynamically according to application requirements with virtualized computational and communication resources. A VG instance is dynamically created for each application and its initial configuration can be further modified during runtime to satisfy the requested level of resource usage and to maintain the quality of service indexes at agreed-upon values. A VG model that takes into account this approach has been proposed and analyzed in the context of Grid-related technology.

Key factors affecting the parallel efficiency of archetypical quantum-chemical calculations are discussed. Effective load balancing schemes are proposed. Introduction of the memory affinity to the balancing process is shown to result in super-linear scaling.

As distributed systems have emerged they become widely adopted by scientific communities and various commercial vendors. Some of them settled architectural models that today are broadly known as grid computing and service oriented architecture. Both of them are base on service and workflow paradigms as a single shareable unit of work. Unfortunately, an increasing number of workflows being brought out has raised a problem of their distribution and management, e.g., search repositories and find workflows similar to the given one in order to increase the efficiency of calculations. In this paper a similar workflow search algorithm based on semantic type comparison has been proposed. In order to evaluate the algorithm usability and precision, an experiment has been conducted that regarded workflow extraction from Feta repository. The entire process involved reasoning based on myGrid ontologies. The received results were compared to results obtained from other algorithm, based on an analysis of the names of the workflow components using the TD-IDF weight. The described experiment shows that semantics and ontology play significant role in service and workflow representation.

The aim of Web service composition is to arrange multiple services into workflows supplying complex user needs. Due to the huge amount of Web services and the need to supply dynamically varying user goals, it is necessary to perform the composition automatically. The objective of this article is to overview the issues of automatic dynamic Web service composition. We discuss the issues related to the semantics of services, which is important for automatic Web service composition. We propose a problem formalization contributing to the formal definition of the pre-/post-conditions, with possible value restrictions, and their relation to the semantics of services. We also provide an overview of several existing approaches dealing with the problem of Web service composition and discuss the current achievements in the field and depict some open research areas.

Intelligent agent-based assistants are systems that try to simplify peoples work based on computers. Recent research on intelligent assistance has presented significant results in several and different situations. Building such a system is a difficult task that requires expertise in numerous artificial intelligence and engineering disciplines. A key point in this kind of system is knowledge handling. The use of ontologies for representing domain knowledge and for supporting reasoning is becoming wide-spread in many areas, including intelligent assistance. In this paper we present how ontologies can be used to support intelligent assistance in a multi-agent system context. We show how ontologies may be spread over the multi-agent system architecture, highlighting their role controlling user interaction and service description. We present in detail an ontology-based conversational interface for personal assistants, showing how to design an ontology for semantic interpretation and how the interpretation process uses it for semantic analysis. We also present how ontologies are used to describe decentralized services based on a multi-agent architecture.

XML documents might often conform to different schemas even in the same application domain. To support the interoperability among different IT systems, this paper proposes a sophisticated method for integrating XML schemas. The proposed method determines the synonym, hypernym, and holonym relationships among XML elements and attributes by using domain ontologies as well as general dictionaries. Specifically, the proposed method takes the structural information of elements and attributes into account. The conciseness of the schema integrated is also considered. Experimental results with a variety of schemas show that the utilization of a domain ontology and the structural information improved the performance of schema integration.