Cláudio R. M. Silva

Optimization of the input impedance of Koch triangular quasi-fractal antennas using genetic algorithms

This work presents an optimization on the input impedance of Koch triangular quasi-fractal antennas using an efficient genetic algorithm (GA). The impedance matching is done by using an inset-fed, which is optimized to minimize the return loss (RL). The excitation of this structure is done using a microstrip line. The antennas are designed using the Ansoft Design™ software and the new structures, that are optimized with GA, are simulated, measured and compared with the same patch antenna but with the lengths of the inset-fed (y0) calculated by well-known models available in the literature. The return loss value of the GA optimized antenna is below -40 dB at the resonant frequency.

A Personality-Based Recommender System for Semantic Searches in Vehicles Sales Portals

This work proposes a personality-based recommender system to implement semantic searches on Internet Vehicles Sales Portals. The system is based on a typical recommender system architecture that has been extended to combine a hybrid recommendation approach with a machine learning classifier technique (k-NN). It proposes a combination of the Five Factor Model (Big Five Model) with a correlation between car fronts and power and sociability perceptions. A prototype was implemented to answer the semantic searches considering personality-based user’s profiles and a set of Brazilian cars. After each search, a questionnaire was provided for the users to verify how successful the recommendations were for them. The prototype received web searches during a period of 15 days. The final report showed that 77.67% of the users accepted the personality-based recommendations, what indicates that the proposed approach could be promising to improve the quality of the recommendations on the user’s point of view.

An Ontology-Based Recommender System Architecture for Semantic Searches in Vehicles Sales Portals

Internet has become an increasingly constant presence everywhere that people go. Particularly this reality is visible in social networks and selling portals scenarios. Whatever scenario, there is plenty of space to improve accuracy since big data is a problem when scale increases. Semantic search is an alternative to improve search accuracy by understanding the contextual meaning of terms as they appear in the searchable data space. Among the several approaches to Semantic Search methodologies, a variation of Ontology-based search (or Logic Approach) is the one adopted. In this methodology, the engine not only understands hierarchical relationships of entities, however also more complex inter-entities relationships defined inside ontologies. This paper proposes a hybrid approach for the problem using Ontology-based Recommender Systems and semantic profiles. A portal prototype is designed and implemented for the domain of online dealerships vehicle buyers market. Precision and Recall measures are the two major indices of information retrieval. They have been used to evaluate the prototype results. After calculating these two metrics over some searches, we have seen that Precision is 86.66% and Recall is 68.42%. These final results have demonstrated an improvement in the searches, particularly with regard the precision of the results provided to the users.

A self-organizing genetic algorithm for UWB microstrip antenna optimization using a machine learning technique

This paper presents an application of a machine learning technique to enhance a multi-objective genetic algorithm to estimate fitness function behaviors from a set of experiments made in laboratory to analyze a microstrip antenna used in ultra-wideband (UWB) wireless devices. These function behaviors are related to three objectives: bandwidth, return loss and central frequency deviation. Each objective (modeled as dependent of an antenna slit dimensions Ls and Ws) is used inside an aggregate adaptive weighted fitness function that estimates the multi-objective behavior in the algorithm. The final results were compared with the ones obtained with a similar antenna modeled in a simulator program and with the ones of a real prototype antenna built from the optimal values obtained after the optimization. The final comparison has shown a promising gain for the designed antenna in the analyzed frequencies.

An alternative approach for particle swarm optimisation using serendipity

In the study of metaheuristic techniques, it is very common to deal with a problem known as premature convergence. This problem is widely studied in swarm intelligence algorithms such as particle swarm optimisation (PSO). Most approaches to the problem consider the generation and/or positioning of individuals in the search space randomly. This paper approaches the issue using the concept of serendipity and its adaptation in this new context. Several strategies that implement serendipity were evaluated in order to develop a PSO variant based on this concept. The results were compared with the traditional PSO considering the quality of the solutions and the ability to find global optimum. The new algorithm was also compared with a PSO variant of the literature. The experiments showed promising results related to the criteria mentioned above, but there is the need for additional adjustments to decrease the runtime.

A Serendipity-Based Approach to Enhance Particle Swarm Optimization Using Scout Particles

In metaheuristic algorithms, such as Particle Swarm Optimization (PSO) and Genetic Algorithm (GA), it is common to deal with a problem known as premature convergence. It happens when a swarm loses diversity and starts converging too early towards a suboptimal solution for an optimization problem. There have been many approaches to this problem along to the latest two decades, but it is a understanding that the problem is still open. This work proposes a new approach based on a concept normally applied in the Recommender Systems context (serendipity-based approach). The paper presents a formalization for the concepts of serendipity and premature convergence, as well a Serendipity-Based PSO (SBPSO) algorithm prototype which implements the concept of serendipity by means of two dimensions: chance and sagacity. The algorithm was compared with the traditional PSO and some PSO variants. The results were successful and showed that SBPSO outperformed the traditional PSO. The experiments also compared SBPSO with some studies in the literature, considering a set of hard functions (such as Rosenbrock, HappyCat, etc) and a fixed number of particles and varying the problem dimensionality and the number of iterations. In all experiments, SBPSO also showed a better convergence behavior, outperforming the traditional PSO and some variants available in the literature regarding the solution quality, the ability to find global optimum, the solutions stability and the ability to restart the movement of the swarm in case of stagnation has been detected.

Uma meta-heurística alternativa de inteligência de enxames baseada em serendipidade guiada

No estudo das tecnicas de meta-heuristica, e muito comum lidar com um problema conhecido como convergencia prematura. Este problema e mais conhecido no contexto dos algoritmos geneticos, mas tem sido observado em outros metodos de meta-heuristica como Particle Swarm Optimization (PSO). A maioria das abordagens para o problema considera a geracao e/ou o posicionamento de individuos no espaco de busca de forma aleatoria. Este trabalho aborda o problema usando o conceito de serendipidade e sua adaptacao neste novo contexto. Varias tecnicas que implementam serendipidade foram avaliadas com o objetivo de construir uma variante do PSO baseada nesse conceito. Os resultados foram comparados com o PSO tradicional e levaram em consideracao a qualidade das solucoes encontradas e a capacidade de localizar otimos globais. O prototipo apresentou resultados promissores com relacao aos criterios citados anteriormente,embora demonstre a necessidade de ajustes adicionais para diminuicao do tempo de execucao.