International Journal of Innovative Computing and Applications (10 papers in press)
Optimal Placement of Unified Power Flow Controller using Differential Search Algorithm
by Dhiman Banerjee, Sriparna Bhattacharya, Provas Kumar Roy
Abstract: Optimal power flow (OPF) problem plays a crucial role to run an economically efficient and well-planned power system. It is a strenuous and challenging task for the power system researchers to cope with the ever-increasing load-demand while getting the minimum system loss. The development of flexible ac transmission system (FACTS) has added a new dimension both to the system operation and research. Unified power flow controller (UPFC) is the most reliable FACTS controller, having its operational capability as series and shunt compensator. As a matter of fact, UPFC can reliably control the different power system parameters. In this article, UPFC is incorporated into the modified IEEE 5-bus and modified IEEE 30-bus test system. Differential search algorithm (DSA) is proposed and implemented to run the OPF with and without UPFC , and the results are listed, analyzed and compared with the same that is obtained by genetic algorithm (GA) and BAT search algorithm.
Keywords: Differential search algorithm; Optimal power flow; FACTS devices; UPFC.
An empirical study of statistical language models: n-gram language models vs. neural network language models
by Freha Mezzoudj, Abdelkader Benyettou
Abstract: Statistical language models are an important module in many areas of successful applications such as speech recognition and machine translation. And n-gram models are basically the state-of-the-art. However, due to sparsity of data, the modelled language cannot be completely represented in the n-gram language model. In fact, if new words appear in the recognition or translation steps, we need to provide a smoothing method to distribute the model probabilities over the unknown values. Recently, neural networks were used to model language based on the idea of projecting words onto a continuous space and performing the probability estimation in this space. In this experimental work, we compare the behaviour of the most popular smoothing methods with statistical n-gram language models and neural network language models in different situations and with different parameters. The language models are trained on two corpora of French and English texts. Good empirical results are obtained by the recurrent neural network language models.
Keywords: language models; n-grams; Kneser-Ney smoothing; modified Kneser-Ney smoothing; Good-Turing smoothing; interpolation; back-off; feed-forward neural networks; continuous space language models; CSLM; recurrent neural networks; RNN; speech recognition; machine translation.
Modelling and implementation of an energy management simulator based on agents using optimised fuzzy rules: application to an electric vehicle
by Rachid El Amrani, Sanaa Faquir, Ali Yahyaouy, Hamid Tairi
Abstract: This paper presents an intelligent algorithm based on multi agent systems to manage the energy in a hybrid electrical vehicle using a model of lithium metal polymer (LMP) battery and a model of an electrical double layer capacitor (EDLC). The algorithm uses fuzzy rules optimised by a genetic algorithm to control the flow of energy inside the system. The LMP battery is linked to a boost converter to insure the autonomy of the electrical vehicle, while the EDLC is linked to a back boost converter that provides the highly demanded energy in a short time and guarantees the temporarily energy storage when the vehicle is braking (no energy is demanded). The hybrid electrical vehicle is simulated in different driving cycles to analyse the behaviour of the LMP battery and the EDLC. Results showed that the used hybrid strategy was able to ensure the autonomy of the vehicle in terms of energy since it has performed a minimum energy cost and a maximum profit in autonomy, which means a longer life of the hybrid electric source.
Keywords: hybrid vehicle; battery; capacitor; modelling; genetic optimisation; fuzzy control; multi-agent system; MAS; energy management.
An algorithm based on Voronoi diagrams for the multi-stream multi-source multicast routing problem
by Romerito Campos De Andrade, Marco César Goldbarg, Elizabeth Ferreira Gouvêa Goldbarg
Abstract: In this study, we present a new heuristic for the multi-stream multi-source multicast routing problem. The core of the heuristic proposed in this study is based on a generalisation of Voronoi diagrams in graphs. It allows building the trees needed to serve the demands of multiple sessions efficiently. Also, the proposed algorithm supports multiple sources. We performed an extensive experimental analysis of different network and problem configurations such as the number of sessions, nodes, sources and participants per session. We compare the proposed algorithm to heuristics proposed previously. The results of the experiments showed that the heuristic proposed in this study finds high-quality solutions efficiently.
Keywords: multicast routing; multi-source; multi-session; Voronoi diagram.
Efficiency analysis of maximum power point tracking techniques for photovoltaic systems under variable conditions
by Rajanand Patnaik Narasipuram, Chaitanya Somu, Ravindranath Tagore Yadlapalli, Lakshmi Sirisha Simhadri
Abstract: With the rapid increase in development of solar energy, researchers are concentrated on developing the maximum power point tracking (MPPT) techniques for extracting the power efficiently. It was environmentally friendly, low maintenance cost, no noise, used in remote areas, and long-lasting life. The output power of PV module is depending on the solar irradiance and temperature. So, to extract more power MPPT techniques are employed. This paper describes the comparative analysis of various MPPT's like perturb and observe (P&O), incremental and conductance (IC), fuzzy logic (FL) controller, and neuro-fuzzy (NF) technique. The controlled output is fed as input to the boost DC-DC converter and the effective performance of the MPPT's are checked under different irradiations and constant temperature. And also, this paper gives mathematical analysis, design and operation of converter and MPPT techniques. In addition, steady state and dynamic performance of MPPT techniques are also analysed. The simulations are performed under MATLAB/Simulink environment.
Keywords: DC-DC converter; photovoltaic module; PV; maximum power point tracking; MPPT; perturb and observe; P&O; incremental and conductance; IC; fuzzy logic controller; FL; neuro-fuzzy; NF.
Special Issue on: Recent Advances in Memetic Computing
Hand Motions Recognition Based on sEMG Nonlinear Feature and Time Domain Feature fusion
by Jiahan Li, Gongfa Li, Ying Sun, Guozhang Jiang, Bo Tao, Shuang Xu
Abstract: In recent years, the development of many rehabilitation robots, bionic prostheses and other sports rehabilitation equipment, which are used to assist the body to restore body movement function, has been paid more and more attention. In the development and design of the current sports rehabilitation equipment or biomimetic prostheses. The classification framework of this paper is a pattern recognition framework. The feature extraction of sEMG is to extract the physical quantity or a set of physical features that fully represent the characteristics of the action class from the electromyogram corresponding to the action of the human hand, in order to distinguish the other types of motion. It's very important step in hand movement recognition. In this paper, the newly developed sEMG nonlinear features AMR are fused with the traditional sEMG time-domain features WL. Feature fusion using SVM-DS fusion algorithm. Hand motions recognition based on feature fusion is improved in accuracy and stability. The accuracy of recognition can be stabilized over 95%.
Keywords: pattern recognition; feature extraction; SVM; D-S evidence theory; Feature fusion; sEMG.
An enhanced cuckoo search using dimension selection
by Lijin Wang
Abstract: This paper proposes an enhanced cuckoo search algorithm using dimension selection. In the proposed strategy, the dimensional distance measure is used to selects a part of dimensions of each solution to search for the new solution in two search components. The dimensions of each solution are selected when those dimensional distances are larger than the average distance of all dimensional distance. A suit of 20 benchmark functions are employed to verify the performance of the proposed algorithm, and the results show the improvement in effectiveness and efficiency of dimension selection.
Keywords: cuckoo search algorithm; dimension selection; dimensional distance; average distance; crossover operator.
A Novel Firefly Algorithm with Self-Adaptive Step Strategy
by Wang Jing
Abstract: In the standard firefly algorithm, the random moving step is very important to the direction of the firefly movement, and the parameter alpha plays an important role in the random moving step. In this paper, we proposed a self-adaptive Step Strategy based on distance control in this paper, and we called it SASFA. Thirteen well-known benchmark functions are used to verify the performance of our proposed method, the computational results show that SASFA is more efficient than many other FA algorithms.
Keywords: firefly algorithm; random moving step; meta-heuristic algorithm; Global optimization;.
A new artificial bee colony based on neighborhood selection
by Xiaoyan Xiong, Jun Tang
Abstract: In this paper, we present a new artificial bee colony (ABC) for solving numerical optimization problems. In the original ABC, a new candidate solution is generated based on the current solution and a randomly selected one. However, the random selection method is unstable. To accelerate the search, a new neighborhood selection is proposed. For each current solution, we firstly randomly select some solutions from the current population. Then, we choose the best one among those solutions as the neighborhood solution to generate new solutions. To verify the performance, we test several classical numerical optimization problems. Simulation results show that our approach outperforms the original ABC and some improved ABC versions.
Keywords: artificial bee colony; neighborhood selection; opposition; global optimization.
One-dimensional deep learning firefly algorithm guided by the best particle
by Zhifeng Xie, Jia Zhao, Hui Sun, Jun Ye, Jiajia Wang, Huasheng Zhu
Abstract: We propose the one-dimensional deep learning firefly algorithm guided by the best particle in order to increase the convergence speed and optimization precision of the firefly algorithm. In each generation of optimization process, the optimal particle is first updated in a fixed number of times according to the newly designed update formula. The update mode is defined as single-dimensional deep learning. After the optimal particle completes single-dimensional deep learning, other fireflies in the population keep the original evolutionary way to update the location and iteratively complete the optimization task. Experiments with 12 benchmark functions show that the proposed algorithm has a higher optimization capacity than the other six modified firefly algorithms.
Keywords: one-dimensional deep learning; firefly algorithm; the optimal particle.