Volumen 7 (2017): Heft 3 (July 2017)

The Modified Condensed Nearest Neighbour (MCNN) algorithm for prototype selection is order-independent, unlike the Condensed Nearest Neighbour (CNN) algorithm. Though MCNN gives better performance, the time requirement is much higher than for CNN. To mitigate this, we propose a distributed approach called Parallel MCNN (pMCNN) which cuts down the time drastically while maintaining good performance. We have proposed two incremental algorithms using MCNN to carry out prototype selection on large and streaming data. The results of these algorithms using MCNN and pMCNN have been compared with an existing algorithm for streaming data.

This contribution presents a numerical evaluation of the impact of parallelization on the performance of an evolutionary algorithm for mixed-integer nonlinear programming (MINLP). On a set of 200 MINLP benchmarks the performance of the MIDACO solver is assessed with gradually increasing parallelization factor from one to three hundred. The results demonstrate that the efficiency of the algorithm can be significantly improved by parallelized function evaluation. Furthermore, the results indicate that the scale-up behaviour on the efficiency resembles a linear nature, which implies that this approach will even be promising for very large parallelization factors. The presented research is especially relevant to CPU-time consuming real-world applications, where only a low number of serial processed function evaluation can be calculated in reasonable time.

Recently, we have witnessed an infusion of calculating models based on models offered by nature, models with more or less fidelity to the original that have led to the development of various problem-solving computational procedures. Starting from the observation of natural processes at the macroscopic or microscopic level, various methods have been developed. Technological progress today allows the accelerated reproduction of natural phenomena in the laboratory, which is why a new niche has arisen in the landscape of nature-inspired methods. This niche is devoted to the emulation of artificial biological processes in computational problem-solving methods.

This paper proposes a novel approach, which is to develop novel computational methods in the field of Natural Computing based on the semi-natural process, namely Directed Evolution. In the first step we explain Directed Evolution, defined as the artificial reproduction of the process of evolution in the laboratory in order to obtain performing biological entities. For computer scientists, this provide a strong source of inspiration in the search for efficient methods of optimization. The computational model that proposed here largely overlaps with the Directed Evolution protocol, and the results obtained in the numerical experiments confirm the viability of such techniques inspired by processes which are more artificial than natural. The paper describes a novel general algorithm, inspired by Directed Evolution, which is able to solve different optimization problems, such as single optimization, multiobjective optimization and combinatorial optimization problems.

Social media is playing an increasingly important role in reporting major events happening in the world. However, detecting events from social media is challenging due to the huge magnitude of the data and the complex semantics of the language being processed. This paper proposes MASEED (MapReduce and Semantics Enabled Event Detection), a novel event detection framework that effectively addresses the following problems: 1) traditional data mining paradigms cannot work for big data; 2) data preprocessing requires significant human efforts; 3) domain knowledge must be gained before the detection; 4) semantic interpretation of events is overlooked; 5) detection scenarios are limited to specific domains. In this work, we overcome these challenges by embedding semantic analysis into temporal analysis for capturing the salient aspects of social media data, and parallelizing the detection of potential events using the MapReduce methodology. We evaluate the performance of our method using real Twitter data. The results will demonstrate the proposed system outperforms most of the state-of-the-art methods in terms of accuracy and efficiency.

The actual motivation of this paper is to develop a functional link between artificial neural network (ANN) with Legendre polynomials and simulated annealing termed as Legendre simulated annealing neural network (LSANN). To demonstrate the applicability, it is employed to study the nonlinear Lane-Emden singular initial value problem that governs the polytropic and isothermal gas spheres. In LSANN, minimization of error is performed by simulated annealing method while Legendre polynomials are used in hidden layer to control the singularity problem. Many illustrative examples of Lane-Emden type are discussed and results are compared with the formerly used algorithms. As well as with accuracy of results and tranquil implementation it provides the numerical solution over the entire finite domain.