Vol. 37 · No. 5

Eventual consistency is demanded nowadays in geo-replicated services that need to be highly scalable and available. According to the CAP constraints, when network partitions may arise, a distributed service should choose between being strongly consistent or being highly available. Since scalable services should be available, a relaxed consistency (while the network is partitioned) is the preferred choice. Eventual consistency is not a common data-centric consistency model, but only a state convergence condition to be added to a relaxed consistency model. There are still several aspects of eventual consistency that have not been analysed in depth in previous works: 1. which are the oldest replication proposals providing eventual consistency, 2. which replica consistency models provide the best basis for building eventually consistent services, 3. which mechanisms should be considered for implementing an eventually consistent service, and 4. which are the best combinations of those mechanisms for achieving different concrete goals. This paper provides some notes on these important topics.

Lately, a lot of research has been conducted on the automatic design of artificial neural networks (ADANNs) using evolutionary algorithms, in the so-called neuro-evolutive algorithms (NEAs). Many of the presented proposals are not biologically inspired and are not able to generate modular, hierarchical and recurrent neural structures, such as those often found in living beings capable of solving intricate survival problems. Bearing in mind the idea that a nervous system's design and organization is a constructive process carried out by genetic information encoded in DNA, this paper proposes a biologically inspired NEA that evolves ANNs using these ideas as computational design techniques. In order to do this, we propose a Lindenmayer System with memory that implements the principles of organization, modularity, repetition (multiple use of the same sub-structure), hierarchy (recursive composition of sub-structures), minimizing the scalability problem of other methods. In our method, the basic neural codification is integrated to a genetic algorithm (GA) that implements the constructive approach found in the evolutionary process, making it closest to biological processes. Thus, the proposed method is a decision-making (DM) process, the fitness function of the NEA rewards economical artificial neural networks (ANNs) that are easily implemented. In other words, the penalty approach implemented through the fitness function automatically rewards the economical ANNs with stronger generalization and extrapolation capacities. Our method was initially tested on a simple, but non-trivial, XOR problem. We also submit our method to two other problems of increasing complexity: time series prediction that represents consumer price index and prediction of the effect of a new drug on breast cancer. In most cases, our NEA outperformed the other methods, delivering the most accurate classification. These superior results are attributed to the improved effectiveness and efficiency of NEA in the decision-making process. The result is an optimized neural network architecture for solving classification problems.

In this paper, we investigate the relationship between the concept lattice and quotient space by granularity. A new framework of knowledge representation – granular quotient space – is constructed and it demonstrates that concept lattice classing is linked to quotient space. The covering of the formal context is firstly given based on this granule, then the granular concept lattice model and its construction are discussed on the sub-context which is formed by the granular classification set. We analyze knowledge reduction and give the description of granular entropy techniques, including some novel formulas. Lastly, a concept lattice constructing algorithm is proposed based on multi-granular feature selection in quotient space. Examples and experiments show that the algorithm can obtain a minimal reduct and is much more efficient than classical incremental concept formation methods.

Text summarization is the task of condensing a document keeping the relevant information. This task integrated in wider information systems can help users to access key information without having to read everything, allowing for a higher efficiency. In this research work, we have developed and evaluated a single-document extractive summarization approach, named SemPCA-Summarizer, which reduces the dimension of a document using Principal Component Analysis technique enriched with semantic information. A concept-sentence matrix is built from the textual input document, and then, PCA is used to identify and rank the relevant concepts, which are used for selecting the most important sentences through different heuristics, thus leading to various types of summaries. The results obtained show that the generated summaries are very competitive, both from a quantitative and a qualitative viewpoint, thus indicating that our proposed approach is appropriate for briefly providing key information, and thus helping to cope with a huge amount of information available in a quicker and efficient manner.

Context-Aware Recommender Systems (CARS) are extensions of traditional recommender systems that use information about the context of the user to improve the recommendation accuracy. Whatever the specific algorithm exploited by the CARS, it can provide high-quality recommendations only after having modeled the user and context aspects. Despite the importance of the data models in CARS, nowadays there is a lack of models and tools to support the modeling and management of the data when developing a new CARS, leaving designers, developers and researchers the work of creating their own models, which can be a hard and time-consuming labor, and often resulting in overspecialized or incomplete models. In this paper, we describe GUMCARS – a General User Model for Context-Aware Recommender Systems, where the main goal is to help designers and researchers when creating a CARS by providing an extensive set of User, Context and Item aspects that covers the information needed by different recommendation domains. To validate GUMCARS, two experiments are performed; first, the completeness and generality of the model are evaluated showing encouraging results as the proposal was able to support most of the information loaded from real-world datasets. Then the structural correctness of the model is assessed, the obtained results strongly suggest that the model is correctly constructed according to Object-Oriented design paradigm.

Cloud computing has become a popular basis that integrated into amount of large platforms to support applications (e.g., multimedia, vehicle traffic, and IoT). It is critical to focus on coordinating the part of these applications that execute in the cloud to provide reliable, scalable and available services. Nevertheless, the problem of optimally coordinating the applications is rarely addressed. In this paper, we develop a stochastic model to analyze the fundamental characteristics that occur in ZooKeeper during the coordination process. The model primarily addresses two aspects: demands of followers and the load of a leader. Then, we derive the optimal strategy for provision with deployment of coordinated servers to achieve load balancing based on various factors (e.g. server capacity and network load), so that the overall network performance is optimized. We evaluate our algorithm under realistic settings and reveal the trend of factors such as CPU, memory utilization and network bandwidth with the increasing number of requests. We propose the algorithm that considers how many servers should be deployed and when. Our results demonstrate that the strategy guarantees the performance by making suitable deployment adjustment.

The quality of recommendations based on any class of recommender systems may become poor if no or low quality data has been provided by users. This is a situation known as it Cold Start problem, which typically happens when a new user registers to the system and no preference data is available for that user. Trust-Aware Recommendation Systems can be considered as a solution for the cold start problem. In these systems, the trust between users plays an import role for making recommendations. However, most of the Trust-Aware RSs consider trust as a context independent phenomenon which means if user a trusts user b to the degree k then user a trusts user b to the degree k in all the concepts. However, in reality, trust is context dependent and user a can trust user b in context X but not in Y. Moreover, most of the trust-aware RSs do not consider an expertise concept for users and all the users are considered as same in the recommendation process. In this paper we proposed a novel approach for detecting expert users just based on their ratings (unlike previous systems which consider the separate profile and extra information for each user to find an expert). In this model a supervised random walk is exploited to search the trust network for finding experts. Empirical experiments on the Epinions dataset shows that EXTRA can outperform previous models in terms of accuracy and coverage.

Cloud computing is nowadays becoming a popular paradigm for the provision of computing infrastructure that enables organizations to achieve financial savings. On the other hand, there are some known obstacles, among which vendor lock-in stands out. Furthermore, due to missing standards and heterogeneities of cloud storage systems, the migration of data to alternative cloud providers is expensive and time-consuming. We propose an approach based on Semantic Web services and AI planning to tackle cloud vendor data lock-in problem. To complete the mentioned task, data structures and data type mapping rules between different types of cloud storage systems are defined. The migration of data among different providers of platform as a service is presented in order to prove the practical applicability of the proposed approach. Additionally, this concept was also applied to software as a service model of cloud computing to perform one-shot data migration from Zoho CRM to Salesforce CRM.

Discrete-Event Systems are discrete in nature, driven by discrete events. Petri Nets are one of the mostly used tools for their modelling and control synthesis. Place/Transitions Petri Nets, Timed Petri Nets, Controlled Petri Nets are suitable when a modelled object is deterministic. When the system model contains uncontrollable/unobservable transitions and unobservable/unmeasurable places or other failures, such kinds of Petri Nets are insufficient for the purpose. In such a case Labelled Petri Nets and/or Interpreted Petri Nets have to be used. Particularities and mutual differences of individual kinds of Petri Nets are pointed out and their applicability to modelling and control of Discrete-Event Systems are described and tested.