Abstract The verification of the Cloud composite services’ correctness is challenging. In fact, multiple component services, derived from different Cloud providers with different service description languages and communication protocols, are involved in the composition which may raise incompatibility issues that in turn lead to a non-consistent composition. In this work, we propose a formal approach to model and verify Cloud composite services. Four verification levels are considered in this article; the structural, semantic, behavioral, and resource allocation levels. Therefore, we opted for the Event-B formal method that enables complex problems decomposition thanks to its refinement capabilities. The proposed approach has proven its efficiency for the modelling and verification of Cloud composite services. The proposed model comprises four abstract levels with respect to the four verification axes. A proof-based approach is applied to the model’s verification. We also succeeded in the validation of the model thanks to the model animation provided by the PROB tool. The use of formal methods provides a rigorous reasoning and mathematical proofs on the correction of the model which ensures the elaboration of correct-by-construction composite services.

Flat routing protocols proposed for MANETs suffer from scalability problem. For this purpose, clustering schemes are proposed to improve the efficiency of routing by organizing the MANET into a hierarchical structure. In this context, we propose SKCA, a stable K-hop clustering algorithm with the view to provide a stable cluster topology and reduce the control overhead. The algorithm proposes a new maintenance function that attempts to reduce cluster topology changes and increases cluster lifetime. Besides, SKCA introduces a novel two-round cluster-head election that reduces the diffusion of the cluster information in the K-hop neighborhood. Using simulation, the performances of SKCA is compared to KCMM algorithm. Also, we study the performances of a cluster-based link state routing protocol in presence of our clustering algorithm SKCA.

Scalability is one of critical challenges for link-state routing protocols in mobile wireless networks. In this context, this paper proposes a proactive link-state protocol based on hierarchical clustering named L-HCLSR. The proposed protocol makes use of clustering algorithm to partition the network and to create a L-levels hierarchical structure of clusters. It uses the structure of clusters to create a virtual backbone useful to broadcast control messages in the entire network. Clustering and virtual structure are used to reduce the routing overhead. Besides, fish eye technique is used to more reduce the routing overhead. We study through simulations the performances of our protocol L-HCLSR with number of level L=1 and L=2, and we compare it to SA-OLSR and F-OLSR protocols. Simulations show that our protocol provides comparable or better performances.

Abductive reasoning (or abduction) is the process of inferring hypotheses from observed data using a certain ‘knowledge’ encoded in the form of inference rules (or causal relations). Many important kinds of intellectual tasks, including medical diagnosis, fault diagnosis, scientific discovery, legal reasoning, and natural language understanding have been characterised as abduction. Unfortunately, abduction is 𝒩𝒫-hard. Genetic algorithms and biologically motivated computational paradigms inspired by the natural evolution turned out to be efficient in solving many hard problems while other existing approaches failed to solve in general. In this article, we present a genetic algorithm called HAKIM, for solving abduction problems. We encode an explanation in a chromosome-like structure, where every gene models a possible single hypothesis. Thereafter, we develop a fitness function that characterises the overall ‘quality’ of a chromosome representing an explanation; and then use standard genetic operators to compute a set of hypotheses that best explains the observed data. Simulation results on large-scale medical problems reveal the good performance of our model HAKIM.

The clustering method “Fuzzy-C-Means” (FCM) is widely used in image segmentation. However, the major drawback of this method is its sensitivity to the noise. In this paper, we propose a variant of this method which aims at resolving this problem. Our approach is based on an adaptive distance which is calculated according to the spatial position of the pixel in the image. The obtained results have shown a significant improvement of our approach performance compared to the standard version of the FCM, especially regarding the robustness face to noise and the accuracy of the edges between regions.

Many real-time applications need data services in distributed environments. Providing such data services is a challeging task due to long remote data accessing delays and stringent time requirements of real-time transactions. In this paper, we use data replication in distributed real-time database systems in order to improve meeting these stringent requirements and fault tolerance . We propose an optimistic replication control protocol, called DLR-ORECOP (Dynamic Level of Replication with Optimistic REplication COntrol Protocol), which finds a trade-off between updating replicas and meeting user transactions deadlines. We introduce a list, called List of Available Copies (LAC), associated with each data item in the database which contains the identifiers of the most updated replicas. Fault tolerance is provided by building LACs dynamically according to transactions executions and system load, giving then the real-time database a dynamic level of replication. The experimental results show that among the replication control protocols evaluated, DLR-ORECOP provides the best performance for a variety of workloads and system configurations. In addition, the experimental results show that DLR-ORECOP performances are not significantly affected by the loss of update messages.

Data mining (DM) is a new emerging discipline that aims to extract knowledge from data using several techniques. DM turned out to be useful in business where the data describing the customers and their transactions is in the order of terabytes. In this paper, we propose an approach for building customer models (said also profiles in the literature) from business data. Our approach is three-step. In the first step, we use fuzzy clustering to categorize customers, i.e., determine groups of customers. A key feature is that the number of groups (or clusters) is computed automatically from data using the partition entropy as a validity criteria. In the second step, we proceed to a dimensionality reduction which aims at keeping for each group of customers only the most informative attributes. For this, we define the information loss to quantify the information degree of an attribute. Hence, and as a result to this second step, we obtain groups of customers each described by a distinct set of attributes. In the third and final step, we use backpropagation neural networks to extract useful knowledge from these groups. Experimental results on real-world data sets reveal a good performance of our approach and should simulate future research.

In this work, we present a framework for the semantic composition of web services based on Statecharts and uniform community service descriptions. Our model is a two step process. In the first step, we derive the execution model of the user's query. The execution model is specified in Statecharts formalism; whereas the user's query is described in OWL-S. Therefore, a mapping from Statecharts formalism to OWL-S is developed. In the second step, we instantiate the developed execution model through invocation of available e-services instances. Hence, and a result, we obtain an execution plan (said also strategy) satisfying user constraints. The key features of the proposed framework could be summarized as follows. First, unlike other existing languages, using OWL-S enables the semantic description of e-services. These semantics are taken into consideration in our composition strategy. Second, the user constraints (or preferences) are taken into account during composition and are expressed as a finite set of logical formulas with the Knowledge Interchange Format (KIF) language.

Clustering is an important research area that has practical applications in many fields. Fuzzy clustering has shown advantages over crisp and probabilistic clustering, especially when there are significant overlaps between clusters. Most analytic fuzzy clustering approaches are derived from Bezdek's fuzzy c-means algorithm. One major factor that influences the determination of appropriate clusters in these approaches is an exponent parameter, called the fuzzifier. To our knowledge, no theoretical reason leading to an optimal setting of this parameter is available. This paper presents the development of an heuristic scheme for determining the fuzzifier. This scheme creates close interactions between the fuzzifier and the data set to be clustered. Experimental results in clustering IRIS data and in code book design required for image compression reveal a good performance of our proposal.

In comparison models for system-level fault diagnosis pairs of units are given the same job and results are compared. The result of such a comparison test can be 0 (match) or 1 (mismatch) and diagnosis is based on the collection of test results. Two such models have been studied, among others: the symmetric model of Chwa and Hakimi and the asymmetric model of Malek. In this paper a novel approach is proposed for identifying faulty units, based on a well-known optimization procedure, as genetic algorithms, which have proven to be useful in various kinds of problems. Furthermore, a new problem-specific genetic mutation is presented and shown to be better than the standard one. A series of simulations was conducted to show the efficiency of the genetic-based approach.

The literature includes a variety of techniques to address the Byzantine Generals Problem or the Byzantine approach. While the main goal within the Byzantine framework is to circumvent and mask the effect of unreliable units (e.g., traitors), much research has been done on identifying (i.e., demasking) unreliable units. This is often called fault identification in system diagnosis. The paper focuses on the identification of unreliable units within the Byzantine framework. Beyond its theoretical interest, an identification of faulty units contributes to accelerating the agreement process itself and drastically reducing the number of messages exchanged between units. The main features of this work are twofold. It does not impose additional assumptions or constraints on the agreement process. It limits the overhead for identifying unreliable units: the identification process is in O(n/sup 3/).

The paper describes a novel approach to the problem of system-level fault diagnosis using genetic algorithms. Consider a system composed of n independent units, each of which tests a subset of the others. It is assumed that at most t of these units are permanently faulty. Such a system is said to be t-diagnosable if, given any complete collection of test results, the set of faulty units can be uniquely identified. Genetic algorithms have recently received much attention as a class of robust stochastic search algorithms for various optimization problems. An efficient method based on evolutionary algorithms is developed to solve the diagnosis problem. The representation of the search space used is in the form of a binary vector of length n. Each bit indicates the status (faulty or fault-free) of its corresponding unit. Genetic operators are adapted to the context of system-level diagnosis. The genetic algorithm was implemented and tested on random test graphs. The simulation results demonstrate the efficiency of the proposed diagnosis algorithm.

Content-based image retrieval is emerging as an important research area with potential applications in many domains, notably multimedia databases and digital libraries. The goal of the present research is to create a usable system that performs automatic picture labeling. Recent studies have shown that some improvements can be achieved with a system which selects the best from amongst many models of representation. The focus of our research is thus twofold: creation of models and selection of the best ones for each label.

In the last decade, abduction has been a very active research area. This has resulted in a variety of models mechanizing abduction, namely within a probabilistic or logical framework. Recently, a few abductive models have been proposed within a neural framework. Unfortunately, these neural/probablistic/logical-based models cannot address complex abduction problems. In this paper, we propose a new extended neural-based model to deal with abduction problems which could be monotonic, open, and incompatible.

In the last decade, abduction has been a very active research area. This results in a variety of models mechanizing abduction, namely within a probabilistic or logical framework. Recently, a few abductive models have been proposed within a neural framework. Unfortunately, these neural-/probabilistic-/logical-based models focus on the most simple class, called independent abduction problems. In this paper, we propose a neural-based model to deal with incompatibility abduction problems. To our knowledge, our model is the first one able to efficiently generate best explanations to the complex class of incompatibility abduction problems.

We address the problem of systematically constructing diagnostic systems for large industrial plants. Toward this end, we propose an environment based on methods, languages and tools allowing systematic construction of diagnostic systems for units of large industrial plants. This environment is based on a skillful articulation of methods, languages, and tools. The process of construction proceeds in three stages as follows. Using methods, the first stage aims at structuring the unit into a set of hierarchies and graphs pointing out several views. The second stage makes use of languages to give a precise specification of the different components embedded in the unit. Finally, tools map the specifications of the unit to derive the diagnostic system.>

Much research has been devoted to diagnosis, where two main approaches have been pointed out: the empirical association-based diagnostic approach and the model-based diagnostic one. Both approaches can be characterized by the kind of knowledge that has to be specified and the diagnostic method that has to be used. However, it seems particularly difficult in real-world applications to obtain a complete description of the faulty (dually, correct) behavior of a system. This incompleteness of description is the reason why deductive reasoning alone is generally insufficient to point out the actual diagnosis. Deduction only allows one to generate some possible partial diagnoses. The latter must be selected and completed to get closer to the actual diagnosis. Both selection and completion require hypothetical reasoning and can be characterized by some preference criteria. The authors' contribution is twofold. A new diagnostic method based on deduction and abduction is then proposed, which is sufficiently flexible to deal with multiple knowledge representations.>

Much research has been devoted to diagnosis, where two main formalizations have been pointed out, The first formalization, called consistency-based diagnosis, stems from the mod- els of correct behavior. The second formalization, called abduction-based diagnosis, considers mainly the models of faulty behavior. Both formalizations contribute to a deep understanding of the diagnostic reasoning process. However, the other major prob- lem in diagnosis is related to the diagnostic modeling task. I. PROLOGUE Much research activities are devoted to diagnosis where different approaches have been explored. Among these approaches let us mention default trees (5), model- based diagnosis (13), associative neural networks (16, 151. From a formal point of view, two formalizations prevail in diagnosis. The consistency-based diagnosis 125, 14, 171 stems from the normality models. This formalization is based on the observation that it is not necessary to de- termine how something, such as a component, is fail- ing to know that it is faulty - a component is faulty if its correct behavior is inconsistent with the observa- tions (17). The abduction-based diagnosis (lo, 19, 91 stems from faulty models. Like human diagnosticians, this formalization takes great advantage of faulty models -known failure knowledge given by experts. A compo- nent is faulty if its observed behavior is consistent with its failure knowledge (17). A thorough survey on these ap- proaches and proposals aiming at their cooperation can be found (13, 4, 23, 17, 31 However, it should be noted that the choice of a faulty or normality model is closely related to the given application domain. For example in medicine it is difficult to make use of normality models since modeling the normal behavior of human body is far from being available. Conversely, for artifacts such as electronic devices it seems attractive (4) to diagnose by using normality models.