International Journal of Bio-Inspired Computation (51 papers in press)
Adaptive Neighborhood Size Adjustment in MOEA/D-DRA
by Meng Xu
Abstract: The multiobjective optimization algorithms based on decomposition(MOEA/D) is a well-known multiobjective optimization algorithms(MOEAs). MOEA/D was proposed by Zhang and Li in 2007s. MOEA/D decomposes a multiobjective problem into a set of scalar single objective subproblems using the aggregation function and the evolutionary operator. The variant of the dynamic resource allocation strategy in MOEA/D(MOEA/D-DRA) has the outstanding performance on CEC2009, the MOEA/D-DRA using the strategy of resource allocation. It cares about the convergence and ignores the diversity. MOEA/D-DRA is very sensitive to the neighbourhood size. In this paper, we present a new enhanced MOEA/D-DRA strategy based on the adaptive neighbourhood size adjustment(MOEA/D-DRA) to increase the diversity. It focuses on the solutions density around of subproblems. The experiment results demonstrate that MOEA/D-ANA strategy performs the best compared with other five classical MOEAs on the CEC2009 test instances.
Keywords: MOEA/D; diversity; Neighborhood; CEC2009 test instances;.
Swarm and Evolutionary Algorithms for Energy Disaggregation: Challenges and Prospects
by Samira Ghorbanpour, Trinadh Pamulapati, Rammohan Mallipeddi
Abstract: Energy disaggregation is defined as the process of estimating the individual electrical appliance energy consumption of a set of appliances in a house from the aggregated measurements taken at a single point or limited points. The energy disaggregation problem can be modelled both as pattern recognition problem and as an optimization problem. Among the two, the pattern recognition problem has been considerably explored while the optimization problem has not been explored to the potential. In literature, researchers have attempted to solve the problem using various optimization algorithms including swarm and evolutionary algorithms. However, the focus on optimization-based methodologies, in general, swarm and evolutionary algorithm based methodologies in particular is minimal. By considering the different problem formulations in the literature, we propose a framework to solve the energy disaggregation problem with swarm and evolutionary algorithms. With the help of simulation results using the existing problem formulations, we discuss the challenges posed by the energy disaggregation to swarm and evolutionary algorithm based methodologies and analyse the prospects of these algorithms for the problem of energy disaggregation with some future directions.
Keywords: Energy Disaggregation; Swarm and Evolutionary Algorithms.
An Improved Brain Storm Optimization Algorithm for Energy-efficient Train Operation Problem
by Boyang Qu, Qian Zhou, Yongsheng Zhu, Jing Liang, Caitong Yue, Yuechao Jiao, Li Yan
Abstract: This paper presents a new method to determine the optimal driving strategies
of the train using an improved brain storm optimization (IBSO) algorithm. In the
proposed method, the idea of successful-parent-selecting frame is applied to improve
the original brain storm optimization (BSO) algorithm avoiding premature convergence
in evolutionary process while dealing with complex problems. The objective of the
algorithm is to minimum energy consumption of the train by nding the switching
points. Furthermore, the speed limits, gradients, maximum acceleration and deceleration
as well as the maximum traction and braking force varying with speed are taken into
consideration to meet practical constraints. Finally the comparison simulations among
four algorithms show that the energy-efficient train operation strategy obtained by IBSO
algorithm are more superior under same conditions.
Keywords: Evolutionary algorithms; Brain Storm Optimization; Energy-efficient train operation.
A path planning method for UAV based on multi-objective pigeon-inspired optimization and differential evolution
by Bingda Tong
Abstract: Inspired by the behavior of pigeon flocks, an improved method of path planning and autonomous formation for unmanned aerial vehicle based on the pigeon-inspired optimization and differential evolution is proposed in this paper. Firstly, the mathematical model for UAV path planning is devised as a multi-objective optimization with three indices, i.e., the length of a path, the sinuosity of a path, and the risk of a path. Then the method integrated by pigeon-inspired optimization and mutation strategies of differential evolution is developed to optimize feasible paths. Besides, Pareto dominance is applied to select global best position of a pigeon. Finally, a series of simulation results compared with standard particle swarm optimization algorithm and standard differential evolution algorithm show the effectiveness of our method.
Keywords: path planning;unmanned aerial vehicle;pigeon-inspired optimization;differential evolution.
Enhanced Pigeon Inspired Optimization Approach for Agile Earth Observation Satellite Scheduling
by Shang Xiang, Lining Xing, Ling Wang, Yongquan Zhou, Guansheng Peng
Abstract: The agile earth observe satellite scheduling problem has been one of the most difficult problem in the field of operations research. For agile satellite, a different observation time means a different observing angle, thus defining a different transition time from the adjacent tasks. Therefore, the agile earth observe satellite scheduling problem has a features of time-dependent which make the solving more difficult. This article formulated a new model of agile earth observe satellite scheduling problem and designed a new solving frame. Enhanced pigeon-inspired optimization was proposed for solving it, in which guide operator and accelerate operator are introduced to enhance the search capability of the algorithm. After test on 15 experimental scenarios, the proposed algorithm own a better performance than the state-of-the-art algorithm.
Keywords: Agile satellite scheduling; pigeon-inspired optimization; guide operator; accelerate operator.
Magnetotactic bacteria optimization algorithm with self-regulation interaction energy
by Jiao Zhao, Hongwei Mo
Abstract: Magnetotactic bacteria optimization algorithm (MBOA) is a new optimization algorithm inspired by the biology characteristics of magnetotactic bacteria in the nature. The original MBOA is vulnerable to premature convergence and has poor exploration capability. In this paper, a modified magnetotactic bacteria optimization algorithm (MMBOA) is proposed to improve the performance of algorithm. It uses a kind of self-regulation interaction energy to enhance the diversity of the swarm for encouraging broader exploration. And as the number of iterations increases, the self-regulation interaction energy can keep MMBOA convergence. Convergence analysis of MMBOA is also implemented. The proposed algorithm converges to the global optimum with a probability one when the number of iterations tends to infinity. Experimental results on CEC2013 and a part of CEC2011 benchmark functions show that the proposed algorithm is efficient and robust, especially in optimizing multimodal functions.
Keywords: Magnetotactic bacteria optimization algorithm; self-regulation interaction energy; exploration; global convergence.
Formation control of multiple UAVs via pigeon swarm optimization
by B.A.I. TINGTING, WANG DAOBO, Zain Anwar Ali, Suhaib Masroor
Abstract: A distributed coordinated control is proposed for the group of UAVs to map the behaviour rule for the pigeon group. Moreover, on the basis of pigeon inspired optimization (PIO), the UAV groups are clustered and synchronized through the coordinated control algorithm. To control the speed, direction of nth UAVs for the desired formation of flight are controlled by proportional integral (PI) and proportional integral differential (PID) controller. A high degree of convergence control method is presented to control the UAV groups at the same level of height during formation of flight. On the basis of pigeon flight obstacle avoidance, a method of UAV obstacle avoidance is designed, and a method of obstacle obstructing is also designed along with the improved artificial physical method. The proposed methodology provides obstacle avoidance for a group of UAV and overcome traditional physical methods, and can easily make the desired formation in which sudden obstacles lie in the way of formation through the arrow channel.
Keywords: Pigeon Swarm Optimization; Unmanned Aerial Vehicle; Pigeon Behavior and Coordinated Control.
An improved NSGA-II with dimension perturbation and density estimation for multi-objective DV-Hop localization algorithm
by Yang Cao, Li Zhou, Fei Xue
Abstract: NSGA-II is a well-known multi-objective optimization algorithm, which has shown excellent performance on many multi-objective optimization problems. However, the classical NSGA-II suffers from uneven distribution of convergence, poor global search ability. To address these issues, this paper proposes an improved NSGA-II (INSGA-II) by employing two strategies: a crossover operation based on dimension perturbation and a novel updating operation based on average individual density estimation. Then the INSGA-II is applied to optimize the multi-objective DV-Hop localization algorithm.
To verify the effectiveness of proposed INSGA-II, we compare it with four other multi-objective evolutionary algorithms on six benchmark functions. Simulation results show that our approach outperforms other compared algorithms. What more, the performance of DV-Hop algorithm based on INSGA-II is tested by the simulation experiments. The simulation results show that DV-Hop localization with INSGA-II achieves better localization accuracy than that with CS, WOCS, MODE and NSGA-II.
Keywords: NSGA-II; dimension perturbation; density estimation; multi-objective optimization; DV-Hop localization algorithm.
Improved Gravitational Search Algorithm Based on Chaotic Local Search
by Zhaolu Guo, Wensheng Zhang, Shenwen Wang
Abstract: The traditional gravitational search algorithm (GSA) maintains good diversity of solutions but often demonstrates weak local search ability. To promote the local search ability of GSA, a new GSA based on chaotic local search (CLSGSA) is introduced in this paper. In its search operations, CLSGSA first executes the conventional search operations of the basic GSA to maintain the diversity of solutions. After that, CLSGSA executes a chaotic local search with the search experience from the current best solution to increase the local search capability. In the experiments, we utilize a suite of benchmark functions to verify the performance of CLSGSA. Moreover, we compare the proposed CLSGSA with several GSA variants. The comparisons validate the effectiveness of CLSGSA.
Keywords: Evolutionary algorithm; optimization algorithm; gravitational search; local search; chaotic map.
A modified single and multi-objective bacteria foraging optimization for the solution of quadratic assignment problem
by Saeid Parvandeh, Mohammadreza Boroumand, Fahimeh Boroumand, Pariya Soltani
Abstract: Non-polynomial hard (NP-hard) problems are challenging because no polynomial time algorithm has yet been discovered to solve them in polynomial time. The Bacteria Foraging Optimization (BFO) algorithm is one of the metaheuristics algorithms that is mostly used for NP-hard problems. BFO is inspired by the behavior of the bacteria foraging such as Escherichia coli (E-coli). The aim of BFO is to eliminate those bacteria that have weak foraging properties and maintain those bacteria that have breakthrough foraging properties toward the optimum. Despite the strength of this algorithm, most of the problems reaching optimal solutions are time-demanding or impossible. In this paper, we modified single objective BFO by adding a mutation operator and multi-objective BFO (MOBFO) by adding mutation and crossover from genetic algorithm operators to update the solutions in each generation, and local tabu search algorithm to reach the local optimum solution. Additionally, we used a fast nondominated sort algorithm in MOBFO to find the best nondominated solutions in each generation. We evaluated the performance of the proposed algorithms through a number of single and multi-objective Quadratic Assignment Problem (QAP) instances. The experimental results show that our approaches outperform some previous optimization algorithms in both convergent and divergent solutions.
Keywords: multi-objective bacteria foraging optimization; genetic algorithm; local tabu search; quadratic assignment problem.
Latent Dactyloscopy Pairing: Presentation Attain Through Feedback from EPITOME
by Pugalenthi R, Mohan Kumar, Arokia Renjith
Abstract: Latent fingerprints play a vital role for recognizing the criminals in the law enforcement agencies. It provides evidence for the court in the law. When the method called AFIS (Automatic Fingerprint Identification System) is applied to rolled and plain fingerprints, it gives a result of high accuracy. Latent accuracy will be less due to the presence of complex background noise, poor ridge quality. In latent the pairing of latent images is a demanding problem. We suggest including top-down data or feedback from an epitome to filter the features taken from a latent for developing latent pairing exactness. The filtered latent features (e.g. ridge orientation and frequency), after feedback, are utilized to re-match the latent to the top X candidate epitome put back by the starting line pairer and retreat the candidate record. The benefactions of this analysis involve: (i) framing ordered steps to utilize data in epitome for latent feature filtration, (ii) improving a feedback pattern which can be delighted around any latent pairer for developing its pairing production, and (iii) deciding when feedback is literally needed to develop latent pairing exactness. Research outcome reveal that combing the proposed feedback pattern with a state-of-the-art latent pairer develops its finding exactness by 0.5-3.5% for NIST SD27 and WVU latent databases against a background database of 100X epitome.
Keywords: Candidate record; Feature filtration; latent fingerprint pairing; epitome feedback.
Expediting Population Diversification in Evolutionary Computation with Quantum Algorithm
by Jun Suk Kim, Chang Wook Ahn
Abstract: Attempts to introduce quantum application to various fields in computer science are growing in numbers as days of commercialized, fully functional quantum computers come closer. Quantum computing\'s uniqueness in commencing parallel computation renders unprecedentedly efficient optimization possible. This paper introduces the adaptation of quantum processing to Crowding, one of the genetic algorithmic procedures to secure undeveloped individual chromosomes in pursuit of diversifying the target population. We argue that the nature of genetic algorithm to find the best solution in the process of optimization can be greatly enhanced by the capability of quantum computing to perform multiple computations in parallel. By introducing the relevant quantum mathematics based on Grover\'s Selection Algorithm and constructing its mechanism in a quantum simulator, we come to conclusion that our proposed approach is valid in such a way that it can precisely reduce the amount of computation query to finish the crowding process without any impairment in the middle of genetic operations.
Keywords: Quantum Computing; Quantum Evolutionary Algorithm; Population Diversification.
Sine-Cosine-Algorithm Based Fractional Order PID Controller Tuning for Multivariable Systems
by Jailsingh Bhookya, Ravi Kumar Jatoth
Abstract: The multivariable systems, used in several industrial processing systems, are controlled using multi-loop controllers. A novel fractional order proportional integral derivative (FOPID) controller design without de-coupler for the multivariable system using an optimization algorithm is proposed. The FOPID controller is employed in each loop of a two-input-two-output (TITO) system and their parameters are tuned using Sine-Cosine-Algorithm (SCA). The proposed method eliminates the interaction of loops with the optimal tuning of FOPID controller parameters. The proposed FOPID design for the Wood-Berry TITO system is compared with the recent state-of-the-art literature to verify its overall efficiency. The simulation result demonstrates the superiority of the proposed controller over other controller designs for the TITO system.
Keywords: Fractional Order PID; Multivariable; TITO System; Multi-loop Control; Sine Cosine Algorithm; Optimal Tuning.
A modified machine learning classification for dental age assessment with effectual ACM-JO based segmentation
by Hemalatha Balan
Abstract: Estimation of Dental age plays a vital role in anthropology, forensics, and bio archaeology. Specific age estimation is mandatory, for living and dead individuals, especially in young adolescents and children. Diverse detection of dental age schemes is calculated, still they have certain limitations such as reliability and prediction accuracy was not superior. To resolve this, Modified Extreme Learning Machine with Sparse Representation Classification (MELM-SRC) is anticipated with dental image in this work. Initially, input image is preprocessed for reducing noise and smoothing in image using Anisotropic Diffusion Filter (ADF). Subsequently, teeth image are segmented using Active Contour Model (ACM) with Jaya Optimization (JO) and then morphological post processing has been applied on segmented result to progress classification accuracy. Next, certain features are extracted such as area, perimeter, solidity, Diameter, major and minor axis length, and filled area to enhance prediction accuracy. Lastly, age has been classified with MELM-SRC. In this MELM, effectual features are classified using SRC to increase age classification accuracy. Simulation outcomes show anticipated MELM-SRC acquires superior performance than Demirjian method for dental age assessment and also other exiting classification schemes such as Radial Basis Function Network (RBFN), and Adaptive Neuro Fuzzy Inference System (ANFIS) schemes.
Keywords: Dental age (DA); Anisotropic Diffusion Filter (ADF); Active Contour Model (ACM); Jaya Optimization (JO) algorithm; Modified Extreme Learning Machine (MELM); Sparse Representation Classification (SRC); Radial Basis Function Network (RBFN); Adaptive Neuro Fuzzy Inference System (ANFIS).
Research on the ensemble feature selection algorithm based on multimodal optimization techniques
by Yan-li Wang, Bo-yang Qu, Jing Liang, Yi Hu, Yun-peng Wei
Abstract: Feature selection is essentially a high-dimensional combinatorial optimization problem. To find representative feature subsets, the selection method needs powerful exploration ability. In addition, if alternative feature subsets could be provided, the final prediction accuracy can be improved by ensembling these subsets. Multimodal optimization (MO) methods with high exploration power and can find multiple suitable solutions in one single run. Therefore, this paper presents an ensemble feature selection algorithm based on multimodal optimization techniques. Differential evolution based on fitness Euclidean-distance ratio (FERDE) algorithm is utilized to search for multiple diverse feature subsets in the huge feature space. A set of diverse base classifiers are built based on these subsets and ensemble to improve the final classification performance. Compared with several existing classical algorithms and ensemble feature selection methods, the proposed method can achieve higher predictive accuracy.
Keywords: feature selection; ensemble learning; multimodal optimization; classification.
A Novel Oversampling Technique Based on the Manifold Distance for Class Imbalance Learning
by Yinan Guo, Botao Jiao, Lingkai Yang, Jian Cheng, Shengxiang Yang, Fengzhen Tang
Abstract: Oversampling is a popular problem-solver for class imbalance learning by generating more minority samples to balance the dataset size of different classes. However, resampling in original space is ineffective for the imbalance datasets with class overlapping or small disjunction. Based on this, a novel oversampling technique based on manifold distance is proposed, in which a new minority sample is produced in terms of the distances among neighbors in manifold space, rather than the Euclidean distance among them. After mapping the original data to its manifold structure, the overlapped majority and minority samples will lie in areas easily being partitioned. In addition, the new samples are generated based on the neighbors locating nearby in manifold space, avoiding the adverse effect of the disjoint minority classes. Following that, an adaptive adjustment method is presented to determine the number of the newly generated minority samples according to the distribution density of the matched-pair data. The experimental results on 48 imbalanced datasets indicate that the proposed oversampling technique has the better classification accuracy.
Keywords: (class imbalance learning; oversampling; manifold learning; overlapping; small disjuncts).
A new approach to design S-Box generation algorithm based on genetic algorithm
by Unal Cavusoglu, Abdullah Hulusi Kokcam
Abstract: Substitution box (S-Box) is one of the most important structures used for byte change operation in block encryption algorithms. An S-Box structure with strong cryptological properties makes the encryption algorithm much more resistant to attacks. In this article, a powerful S-Box generation algorithm (GA) design is presented using genetic algorithm. In the genetic algorithm based S-Box generation algorithm, the nonlinearity value which is one of the most important of the S-Box evaluation criteria, has been processed. Quality of the generated S-Boxes is determined by performance tests. Obtained performance results are compared with the S-Boxes in the literature. It has been found that the presented algorithm generates S-Boxes with strong cryptological properties.
Keywords: Genetic Algorithms; Information Security; Nonlinearity; S-Box.
Cauchy-Gaussian pigeon-inspired optimization for electromagnetic inverse problem
by Mengzhen Huo, Yimin Deng, Haibin Duan
Abstract: The optimization of electromagnetic inverse problems could be attributed to a constraint nonlinear programming problem. Loneys solenoid problem is one of the electromagnetic inverse benchmarks in the magnetic field. Parameters such as the structure and medium are necessary to be designed based on the required magnetic properties. In this paper, an improved variant of pigeon-inspired optimization (PIO) algorithm based on Cauchy distribution and Gaussian distribution, named Cauchy-Gaussian pigeon-inspired optimization (CGPIO), is proposed to solve the electromagnetic inverse problems. The PIO algorithm is a bio-inspired swarm intelligence optimization algorithm, which imitates the homing process of pigeons. To improve the convergence efficiency of the basic PIO algorithm, two operators including Cauchy distribution and Gaussian distribution are utilized. Comparative results show the suitability and superiority of CGPIO algorithm for electromagnetic optimization.
Keywords: Pigeon-inspired optimization; electromagnetic inverse problem; Loney’s solenoid problem; Cauchy distribution; Gaussian distribution.
A real adjacency matrix-coded evolution algorithm for highly linkage-based routing problems.
by Hang Wei, Han Huang, Zhi-Feng Hao, Witold Pedrycz, Qin-Qun Chen, Gang Li
Abstract: In routing problems, the contribution of a variable to ?tness often depends on the states of other variables. This phenomenon is referred to as linkage. High linkage level typically makes a routing problem more challenging for an evolutionary algorithm (EA). An entire linkage measure, named entire linkage index (ELI), has been proposed in this paper for such routing problems. Aiming at solving high linkage-based routing problems, we presented a real adjacency matrix-coded evolution algorithm (RAMEA) that is capable of learning and evolving correlation matrix of decision variables. The e?ciency of RAMEA was tested on two familiar routing problems: traveling salesman problem (TSP) and generalized traveling salesman problem (GTSP). The experimental results show that the RAMEA is promising for those highly linkage-based routing problems, especially for those of large scale.
Keywords: linkage; routing problems; evolutionary optimization; real adjacency matrix-coding mechanism.
Comparison of Three Nature Inspired Molecular Docking Algorithms
by Petra Cechová, Martin Kubala
Abstract: Molecular docking uses different methods to generate and evaluate the binding between a receptor and a ligand. Three nature-inspired docking programs are compared on a test set of 65 receptor-ligand pairs. AutoDock uses a genetic algorithm inspired by the process of natural selection; PSOVinaLS uses the Particle Swarm Optimization method, based on the behaviour of animal flocks and PLANTS uses an algorithm based on ant colonies. Using the default parameters, PSOVinaLS achieved the best performance with respect to time required and docking accuracy, followed by PLANTS and, with a large gap, AutoDock. However, all the programs exhibited difficulties with redocking of ligands with more than 10 rotable bonds.
Keywords: Molecular docking; redocking study; genetic algorithm; particle swarm optimization; ant colony optimization;.
An Integrated Data Mining Approach to Predict Electrical Energy Consumption
by Alireza Fallahpour, Kaveh Barri, Kuan Yew Wong, Pengcheng Jiao, Amir H. Alavi
Abstract: This study proposes an integrated adaptive neuro fuzzy inference system (ANFIS) and gene expression programming (GEP) approach to predict long-term electrical energy consumption. The developed hybrid method uses ANFIS to find parameters with maximum effect on the electricity demand. Thereafter, the GEP algorithm is deployed to derive a robust mathematical model for the prediction of the electricity demand. Various statistical criteria are considered to verify the validity of the model. The predictions made by the ANFIS-GEP model are compared with those obtained by the simple GEP and hybrid artificial neural network (ANN)-ANFIS methods. The proposed ANFIS-GEP technique is more computationally efficient and accurate than GEP, and notably outperforms ANFIS-ANN.
Keywords: Electricity demand forecasting; Feature selection; ANFIS; GEP; Formulation.
A New Hybrid System Combining Active Learning and Particle Swarm Optimization for Medical Data Classification
by Nawel Zemmal, Nabiha Azizi, Mokhtar Sellami, Soraya Cheriguene, Amel Ziani
Abstract: With the increase of unlabeled data in medical datasets, the labelling process becomes a more costly task. Therefore, Active Learning provides a framework to reduce the amount the manual labor process by querying an expert for just the labels of particular instances, the choice of these instances to annotate is paramount. However, the traditional active learning techniques can be computationally expensive as they require to analyze, at each iteration, all unlabeled instances including those that are redundant and uninformative, thereby decreasing the system performance. To handle this issue, it is necessary to have a global optimization algorithm that allows finding the best solution in a reasonable time. This paper proposes a novel framework combining active learning and particle swarm optimization algorithm. A novel uncertainty-based strategy was designed and integrated into the PSO as an objective function. This new strategy allows finding the most informative instances by calculating an uncertainty score using instance weighting method. Experiments were performed on binary and multi-class classification problems using both balanced and unbalanced medical datasets. Experimental results show that the proposed uncertainty strategy outperforms its existing counterparts. It achieves performances comparable to supervised methods.
Keywords: Active Learning; Uncertainty Sampling Strategy; Particle Swarm Optimization; Global Optimization Problem; Instance Weighting.
Density peaks clustering based on geodetic distance and dynamic neighborhood
by Li Lv, Jiayuan Wang, Runxiu Wu, Hui Wang, Ivan Lee
Abstract: Density peaks clustering algorithm uses Euclidean distance as a measure of similarity between the samples, and it can achieve a good clustering effect when processing the manifold datasets. Utilizing this feature, we propose a density peaks clustering algorithm based on geodetic distance and dynamic neighborhood. This new algorithm measures the similarity between the samples by using geodetic distance, and the number of neighbors K is dynamically adjusted according to the spatial distribution of samples for geodetic distance computation. By choosing geodetic distance as the similarity measure, the problems of manifold dataset clustering can be easily solved, and the clustering is made more effective when the sparse clusters and dense clusters co-exist. The new algorithm was then compared against the other 5 clustering algorithms on 6 synthetic datasets and 10 real-world datasets. The experiments showed that the proposed algorithm not only outperformed the other conventional algorithms on manifold datasets, but also achieved a very good clustering effect on multi-scale, cluttered and intertwined datasets.
Keywords: Density peaks; clustering; geodetic distance; dynamic neighborhood.
DVR Based Power Quality Enhancement Using Adaptive Particle Swarm Optimization (APSO) Technique
by Ravi Srinivas Lanka, Mahesh Babu Basam, Tulasi Ram S.S.
Abstract: This paper proposes a heuristic control of the series active power filter for power quality enhancement. In the context of power quality, the series active filter is better utilized as a voltage source controller contrary to its conventional usage as variable impedance. The present-day utility system as a linear model is unsatisfactory and the steps are laid down to discuss utility systems as a nonlinear model. This paper deals with comparative analysis of particle swarm optimization (PSO) and its contemporary modified PSO algorithms modified PSO (MPSO), Adaptive PSO (APSO). The harmonic reduction in the source current and sags/swells mitigation in the load voltage is carried out with the optimal tuning of the PI controller. The Series active power filter as a harmonic suppressor with a specific reference controlled strategy is discussed in this paper. The synchronous reference frame (SRF) Theory and its modified versions are used to generate a compensating signal. The hysteresis band current controller (HBCC) is used to perform the switching operation of Voltage Source Inverter. Simulations are carried out in the MATLAB/SIMULINK environment.
Keywords: Dynamic Voltage Restorer (DVR); Synchronous Reference Frame theory (SRF); Particle Swarm Optimization (PSO); Modified PSO (MPSO); Adaptive PSO (APSO).
Pigeon-inspired optimization algorithm with hierarchical topology and receding horizon control for multi-UAV formation
by Yankai Shen, Yiming Deng
Abstract: Multi-UAV formation is a crucial research aspect of UAV cooperation control. This paper proposes a hierarchical topology pigeon-inspired optimization (HPIO) and a receding horizon control (RHC) to deal with the problem. The RHC method can convert the problem into an online optimization problem. Besides, inspired by the structure of pigeon flock, a new topology with a hierarchical structure of pigeons is designed. The velocity updating formula of the original PIO is redesigned. Then discuss the stability of the modified PIO. Finally, use the HPIO to generate the desired formation control in each time domain of the RHC. The simulation results verify the feasibility and effectiveness of the proposed method.
Keywords: Multi-UAV Formation; Receding horizon control (RHC); Pigeon-inspired optimization (PIO); Hierarchical topology.
Top-Down modulated model for object recognition in different categorization levels
by Fatemeh Sharifizadeh, Mohammad Ganjtabesh, Abbas Nowzari-Dalini
Abstract: The human visual system contains a hierarchical sequence of modules that take part in visual perception at superordinate, basic, and subordinate categorization levels. The top-down signals facilitate the bottom-up processing of visual information in the cortical analysis of object recognition. We propose a novel computational model for object recognition in different categorization levels, which mimics the effects of top- down signals in the hierarchical processing of the visual system. The top-down signal is incorporated in bottom-up processing of input image to increase the biological plausibility of our model as well as its efficiency for the object recognition in different categorization levels. The top-down signals provide a pre-knowledge about the input space, which can help to solve the complex object recognition tasks. The performance of our model is evaluated by various appraisal criteria with three benchmark datasets and significant improvement in recognition accuracy of our proposed model is achieved in all experiments.
Keywords: Object Recognition; Categorization Levels; Computational Models; Bottom-Up Processing; Top-Down Signals.
Inspiration-wise Swarm Intelligence Meta-heuristics for Continuous Optimization: A Survey--- Part III
by Nadia Nedjah, Luiza De Macedo Mourelle Mourelle, Reinaldo Gomes Moraes
Abstract: Optimization is an intrinsic part of many sciences such as engineering, logistics, transportation and economics among others. Global optimization techniques aim to find the best solution of a given a problem. In general, these problems are complex, nonlinear and intractable. On the other hand, there is a plethora of optimization techniques. Meta-heuristics are general optimization efficient tools. Swarm Intelligence is based on models inspired by swarming behaviors. The literature bears many swarm based meta-heuristics. In the first part of this survey, we propose an inspiration-wise taxonomy of such meta-heuristics and reviewed meta-heuristics that are guided by known human relation's characteristics and those that are inspired by physical system's properties. In the second part, we surveyed existing techniques that are inspired by fictional metaphors as well as flocking or schooling behaviors, which are considered to be bio-inspired. In this third part of the survey, we concentrate on bio-inspired methods that are guided by swarming, herding and proliferating behaviors. As an overall result of the survey, we point out the common inconvenience of using many of these meta-heuristics, which is related to the setting of the underlying parameters of the search strategy. Needless to state that their performance is dependent on the setting of these parameters. Nonetheless, there are some proposed techniques that reduce the required parameters to a minimum. Moreover, dynamic adjustment of these parameters during the search process is usually exploited to mitigate the impact of the parameter calibration process.
Keywords: Swarm Intelligence; Swarm-based Meta-heuristics; Bio-inspired Computation.
Classification of Diversified Web Crawler Accesses Inspired by Biological Adaptation
by Naomi Kuze, Shu Ishikura, Takeshi Yagi, Daiki Chiba, Masayuki Murata
Abstract: To discover and prevent attacks, it is necessary to collect data about the attacks using honeypots, and to identify malicious accesses from collected data.
In this study, we focus on detecting a massive number of crawler accesses, which complicates the detection of malicious accesses.
We adapt AntTree, a bio-inspired clustering scheme that is highly scalable and adaptable, for crawler detection.
We also desgin a feature vector for crawler detection and propose a cluster interpretation method of AntTree.
Our results show that the proposed bio-inspired mechanism can detect crawlers with a low false-negative rate, which is an advantage over conventional schemes for detecting various types of crawler.
Keywords: Network security; web vulnerability scanning detection; web honeypots; ant-based clustering.
Research on Interface Design Based on User\'s Mental Model Driven by Interactive Genetic Algorithm
by Zhen Wei, Jinghuan Nie
Abstract: This study proposes a computer interactive design method based on the analytic hierarchy process (AHP) and interactive genetic algorithm (IGA) to explore the mental model of user interface and promote the optimal design of interface. IGA is used in this study to display to users directly in the form of visual coding, and the most satisfied visual combination of users is obtained by iteration of user rating method. At the same time, in order to alleviate the problem of user fatigue and the problem that some design factors are not easy to be visual coded in IGA, AHP is introduced to obtain design factors user satisfied that are not easy to be visual coded, and the user preference is obtained by sorting the weight value of factors, then the user mental model is formed and the user mental model is transformed into design elements and integrate into visual design.
Keywords: user mental model; interactive genetic algorithm; analytic hierarchy process; interactive interface design.
Pigeon inspired optimization based cooperative target searching for multi-UAV in uncertain environment
by Delin Luo
Abstract: In this paper, the multi-UAV cooperative target searching problem is investigated and a close loop path planning method is developed for UAVs in uncertain environment. The proposed method includes two consecutive parts, the multi-UAV cooperative target search algorithm developed based on cooperative pigeon inspired optimization (CPIO) and the base returning algorithm for each UAV based on artificial potential field (APF) method. Firstly, a concerned regional environment and the initial search probability map models are established. Then, by applying the rolling prediction strategy, the cooperative target search paths for multiple UAVs are generated by utilizing the proposed CPIO. With this method, UAVs can reinforce target search in the key areas in a cooperative way and avoid flying into the no-fly zones. In the meanwhile, the Bayesian theorem is used to constantly update the search probability map in each search step. Finally, at the end of the target search phase, an optimized safe path is generated for each UAV returning back to its original by using the APF method. Simulations are performed and the results demonstrate that the proposed approach is effective for multiple UAVs carrying out cooperative target search task in a complex environment.
Keywords: multi-UAV; cooperative search; pigeon-inspired algorithm; artificial potential field.
Binary Fireworks Algorithm Application for Optimal Schedule of Electric Vehicle Reserve in Traditional and Restructured Electricity Markets
by Srikanth Reddy, Lokesh Panwar, B.K. Panigrahi, Rajesh Kumar
Abstract: The electric vehicle (EV) is a proven and reliable technology to curtail global emissions in the transportation sector. The multipurpose use of EV makes it a special entity not only in transportation sector but also in the modern smart grid. This paper presents an application of Binary Fireworks Algorithm for solving a multi-beneficiary model in which conventional thermal units are scheduled considering EV aggregator as an identical entity to GENCO in offering reserve services. The fast response of EV can be exploited in providing reserve capacity along with traditional unit commitment (UC) in regulated and price based unit commitment (PBUC) of thermal generators/units in deregulated electricity markets respectively. This paper extends the application of Binary Fireworks Algorithm (BFWA) to resource scheduling in smart grid to optimize the generation schedule of thermal and EV reserve in the day-ahead market. Also, develops first application of BFWA algorithm to solve PBUC. Thereafter, the resource scheduling objective in PBUC is extended to include EV reserve. Simulation results are presented for discussion and the same reveals that such reserve market participation of EV can improve the economics of both utility and EV. Also, compared to energy market participation, in reserve market profit for EV is increased by $3.59/MW and $3.96/MW in deregulated and traditional market scenario respectively. In addition, comparative analysis for two types of EV namely, battery electric vehicle (BEV), fuel cell electric vehicle (FCEV) is carried out with respect to cost-benefits. The market participation, especially reserve market participation proved much profitable for BEV compared to FCEV.
Keywords: Binary fireworks algorithm; Computational Intelligence; Unit Commitment; Profit Based unit Commitment (PBUC); Electric vehicle (EV); Electricity markets.
3D Reconstruction of Structural Magnetic Resonance Neuroimaging based on Computer Aided Detection
by Mai S. Mabrouk, Heba M. Afifi, Samir Y. Marzouk
Abstract: Computer Aided Detection (CAD) has numerous achievements in medical image processing field which represents a communication system between human recorded brain images and software for brain detection and diagnosis. CAD technology for brain images is the integration between scientists, engineers, and clinicians for analyzing brain images to help patients suffered from neuromuscular disorders, brain tumors, cerebral palsy, stroke, spinal cord injury, Epilepsy, Alzheimer and Parkinsons diseases. Magnetic resonance imaging (MRI) is an extensively used tool for discovery of neural diseases and determination of normal and abnormal brain images to support the radiologists without the need for surgical biopsy or resection. Despite the advance in the radiological diagnosis of neuroimaging, MRI has some restrictions related to human errors and incomplete interpretation of brain tumor regions. Also, MRI produces 2D images of the brain that was very difficult to handle different types of tumor. Therefore, computer-based classification is used to accurately distinguish between tumor regions from the brain MR images that provided early diagnosis and medical decision-making operation for brain diseases. This study investigated CAD system using 3D image reconstruction of MR brain and tumor structures efficiently. In addition, the proposed system applied the Fuzzy C-Means (FCM) algorithm as image segmentation and support vector machine (SVM) as image classification for tumor detection of MR brain images. Results confirmed that this 3D model supported the advanced view of human brain diseases. Besides, accurate and quantitative research results are expected for enhancing the healthcare technology. This model will open a new framework for CAD technology to restore 3D brain images
Keywords: Computer Aided Detection (CAD); Magnetic resonance imaging (MRI); 3D image reconstruction; support vector machine (SVM); Fuzzy C-Means (FCM) algorithm.
An improved extension neural network methodology for fault diagnosis of complex electromechanical system
by Yunfei Zhou, Yunxiu Sai, Li Yan
Abstract: Fault diagnosis is a complex and challenging problem in the operation of machines and equipment, the crucial component of which is the recognition, extraction and classification of fault features, which are closely related to the training sets. A good diagnostic method can yield twice the result with half the effort result. In the existing intelligent diagnostic methods, artificial neural network (ANN) is widely used to identify and classify fault patterns. Although this approach solves the Fault diagnosis problem of complex electromechanical system (CES), there are still two defects: (1) The training process converges slowly; (2) Feature recognition and extraction are largely influenced by the training data sets. In order to solve the above problem, we propose a novel intelligent fault diagnosis method based on the extension neural network (ENN) and a uniform distribution search for particle swarm optimization (UPSO). The data set of the turbine generator set verifies the validity of this method. The diagnostic results show that compared with the existing method, the proposed method can not only extract the available classical domain characteristic from the collected training data set adaptively, but also have high diagnostic accuracy and rapid training process.
Keywords: fault diagnosis; complex electromechanical system; particle swarm optimization; extension neural network.
ELM-NeuralWalk: Trust evaluation for online social networks
by Shuoshuo Zhang, Xiangrong Tong, Shuigen Wang
Abstract: Trust relationship plays an important role in online shopping, recommendation systems, Internet of Things, etc. The problem of trust evaluation among users in online social network (OSN) has attracted much attention, and has become a hot issue in the domain of social computing. However, the way of trust propagation and aggregation in OSN is still not clear, as well as the accuracy of trust calculation. In order to calculate the indirect trust, an ELM-NeuralWalk algorithm to implement trust propagation and aggregation is proposed. ELM-WalkNet firstly learns two-hop trust calculation rules, calculates two-hop trust among users in the OSN. After that, ELM-NeuralWalk updates the OSN with the calculated trust value, so as to realize the calculation of multi-hop trust among users through iterative calling ELM-WalkNet. Unlike traditional solutions that use inference methods, ELM-WalkNet can learn trust calculation rules in an inductive way and accurately calculate indirect trust between users. Experiments on two real OSN datasets showed that ELM-NeuralWalk outperforms existing solutions.
Keywords: Trust; online social networks; extreme learning machine; machine learning.
Learning the Number of Filters in Convolutional Neural Networks
by Jue Li, Feng Cao, Honghong Cheng, Yuhua Qian
Abstract: Convolutional networks bring the performance of many computer vision tasks to unprecedented heights, but at the cost of enormous computation load. To reduce this
cost, many model compression tasks have been proposed by eliminating insignificant
model structures. For example, convolution filters with small absolute weights are pruned and then fine-tuned to restore reasonable accuracy. However, most of these works rely on pre-trained models without specific analysis of the changes in filters during the training process, resulting in sizable model retraining costs. Different from previous works, we interpret the change of filter behavior during training from the associated angle, and propose a novel filter pruning method utilizing the change rule, which can remove filters with similar functions later in training. According to this strategy, not only can we achieve model compression without fine-tuning, but we can also find a novel perspective to interpret the changing behavior of the filter during training. Moreover, our approach has been proved to be effective for many advanced CNN architectures.
Keywords: model compress; filter pruning; filter correlation; filter behavior interpretable.
Jumping particle swarm optimization method for solving minimum weight vertex cover problem
by Balaji Sankarasubramanian, Kanagasabapathy G, Vikram S. T.
Abstract: The minimum weight vertex cover problem (MWVCP) is one of the classic combinatorial optimization problem. In this paper we have proposed a modified jumping particle swarm optimization approach, named JPS-VC, for MWVCP. This method contains two new local search approaches incorporated with the jumping particle swarm optimization approach and it admits several remarkable features such as novel construction of feasible solution and also the quick conversion of feasible solution to optimal solution. The computed results indicate that the proposed approach has got appreciable performance over the recent state-of-the-art algorithms.
Keywords: minimum weight vertex cover; discrete particle swarm optimization; computational methods; combinatorial optimization.
Effect of potential well model for quantum heuristic algorithm: A comparative study and application
by Jin Jin, Peng Wang
Abstract: The multiscale quantum harmonic-oscillator algorithm (MQHOA) is an intelligent optimization algorithm based on quantum harmonic wave functions. Although it is effective for many optimization problems, an analysis for its performance is still lacking. This paper discusses the harmonic-oscillator potential well, delta-function potential well, and infinite-square potential well in terms of their application in evolutionary algorithms. Of the three, the harmonic-oscillator potential well is considered to give the most precise approximation for complex objective functions. To verify its global optimization performance, experiments are conducted using a suite of benchmark functions to compare the performance of different potential wells and heuristic algorithms. The experimental results indicate that MQHOA with the harmonic-oscillator potential well is a better practical choice than the other two potential well models, and show that MQHOA is a potential quantum heuristic algorithm.
Keywords: Multiscale quantum harmonic oscillator algorithm; quantum heuristic algorithm; global optimization; potential well; wave function.
A novel whale optimization algorithm with filtering disturbance and non-linear step
by Jinkun Luo, Fazhi He, Haoran Li, Xian-Tao Zeng, Yaqian Liang
Abstract: As a recent addition of population-based meta-heuristic algorithms, whale optimization algorithm (WOA) has attracted a lot of attention recently. However, WOA still has room for improvement in the accuracy and reliability of solution. In this study, a novel WOA with filtering disturbance and non-linear step (FDNS-WOA) is proposed. Firstly, to enhance the population diversity and the global search ability, we design a weighted Cauchy mutation equation to perturb solution space. Secondly, the excessive disturbance may cause the imbalance between exploration and exploitation, and the quality of solutions will be decreased. Thus, before the disturbance, we propose a filtering mechanism to dynamically select different individuals for disturbance at different evolutionary stages. Thirdly, a weighted local walk mechanism is designed to improve the local exploitation capability. Different from original linear step with large fluctuation, we propose a non-linear step to reduce blind spots in the local exploitation process. Finally, the proposed FDNS-WOA is tested on benchmark functions and applied in a real world problem. The experimental results show that the proposed FDNS-WOA not only outperforms other recent meta-heuristic algorithms in term of accuracy and reliability for most benchmark functions, but also achieves satisfactory results in real world problem.
Keywords: whale optimization algorithm; Cauchy mutation; economic load dispatch; meta-heuristic algorithm; population-based algorithm; filtering disturbance; non-linear step.
An adaptive multi-objective particle swarm optimization algorithm based on fitness distance to streamline repository
by Suyu Wang, Dengcheng Ma, Ze Ren, Yuanyuan Qu, Miao Wu
Abstract: In recent years, multi-objective particle swarm optimization algorithm (MOPSO) has been paid more attention. One of its indispensable structures is the maintenance and update mechanism of the repository. The existing mechanisms are relatively simple, and most of them are based on the crowding distance sorting strategy, and not conducive to the distribution and accuracy of the algorithms. The paper innovated this mechanism and proposed an adaptive multi-objective particle swarm optimization algorithm to streamline repository based on fitness distance (FDMOPSO). Both the concept of fitness distance and the corresponding improve methods of mutation mechanism and adaptive mechanism were proposed. The algorithm itself was tested using benchmarks. The results shown that the proposed application of fitness distance had a better improvement on the convergence and distribution. Compared with other algorithms, the FDMOPSO algorithm had the best overall performance.
Keywords: MOPSO; fitness distance; streamline repository; adaptive; multi-objective optimization.
Efficient Optimization Methods in a Distributed Memory Storage System Using Data Compression
by Xiaoyang Yu, Songfu Lu, Tongyang Wang, Xinfang Zhang, Shaohua Wan
Abstract: The storage of data in a memory-centric storage system benefits from great improvements in read and write I/Ornperformance at the price of consuming a large amount of valuable memory resources. As the amount of data grows, a distributedrnsystem must scale accordingly, resulting in an increased number of nodes and greater communication overhead that may impact thernoverall performance. We believe that achieving an appropriate trade-off among computation, storage and network transportation will bernbeneficial for a distributed memory storage system, leading to higher overall efficiency. The challenges of achieving such a trade-off arerndetermining how and when this trade-off should be applied. In this paper, we explore a method to achieve this trade-off by introducingrndata compression technology in a transparent manner. Instead of focusing on specific compressed data structures, we target blockrnlevel compression for a general purpose storage system to incorporate a wide range of existing data analysis frameworks and usagernscenarios, especially with big data. A prototype is implemented and evaluated based on the memory-centric distributed storage systemrnAlluxio to provide transparent compression and decompression during write/read operations. A series of experiments for data withrndifferent types of compression ratio is conducted to determine the proper situation in which this technology can help to improve thernsystem efficiency. The experimental results prove that our approach can achieve huge write/read throughput boosts for highlyrncompressible data and performs better for moderately compressible data when there are remote operations. From the results, wernsuggest that our approach can be applied for data with a medium up to high compression ratio in practical data processing pipelinesrnand provides useful insights for optimizing future memory-based storage systems.
Keywords: System performance optimization; Data compression; Distributed memory storage.
Join query optimization in the distributed databases using a hybrid harmony search and artificial bee colony algorithm
by Manizhe Yazdani, Nima Jafari Navimipour
Abstract: The query optimizer is an important part of any distributed database system. Reducing the execution time of query depends on finding an optimal query execution plan. Due to the NP-hard nature of this issue, a hybrid harmony search and artificial bee colony algorithm can provide a new method in the distributed databases. The harmony is used to call each query plan and represented by an S-dimension real vector. A harmony memory is a place for creating and storing a primary population of harmony vectors. Then, bees explore and exploit the harmony memory as food sources. The production of new candidate harmony from all of the query plans in the harmony memory requires a pitch adjustment rule, a memory consideration rule, and a random re-initialization. Finally, the new candidate vector replaces the worst harmony vector when it works better. The simulation results have shown that the proposed algorithm reduces the cost of evaluating a query compared to harmony search algorithm and bee colony optimization, but this method has longer execution time.
Keywords: Distributed database system; Join query optimization; Harmony search algorithm; Artificial bee colony.
Research on multi-UAV task decision-making based on improved MADDPG algorithm and transfer learning
by Bo Li, Shiyang Liang, Zhigang Gan, Daqing Chen, Peixin Gao
Abstract: At present, the intelligent algorithms of multi-UAV task decision-making have been suffering some major issues, such as, slow learning speed and poor generalization capability, and these issues have made it difficult to obtain expected learning results within a reasonable time and to apply a trained model in a new environment. To address these problems, an improved algorithm, namely PMADDPG, based on Multi-Agent Deep Deterministic Policy Gradient (MADDPG) is proposed in this paper. This algorithm adopts a two-layer experience pool structure in order to achieve the priority experience replay. Experiences are stored in an experience pool of the first layer, and then, experiences more conducive to training and learning are selected according to priority criteria and put into an experience pool of the second layer. Furthermore, the experiences from the experience pool of the second layer are selected for model training based on PMADDPG algorithm. In addition, a model-based environment transfer learning method is designed to improve the generalization capability of the algorithm. Comparative experiments have shown that, compared with MADDPG algorithm, proposed algorithms can scientifically improve the learning speed, task success rate and generalization capability.
Keywords: Multi-UAV task decision; Improved MADDPG algorithm; Two-layer experience pool; Transfer learning.
A new multi-objective artificial bee colony algorithm based on reference point and opposition
by Songyi Xiao, Wenjun Wang, Hui Wang, Zhikai Huang
Abstract: A new multi-objective artificial bee colony (ABC) algorithm based on reference point and opposition (called ROMOABC) is proposed in this paper. Firstly, the original framework of ABC is modified to improve the efficiency of population renewal and accelerate the convergence rate. On basis of this framework, two new strategies are proposed. In the scout bee search, opposition-based learning and elite solutions are used to reduce the waste of computing resources. Distribution of solutions is improved by using reference points associated external archive. Experiments are conducted on sixteen multi-objective benchmark functions including ZDT, DTLZ and WFG multi-objective benchmark functions. The comparison of ROMOABC with five other multi-objective algorithms shows that it has competitive convergence and diversity.
Keywords: artificial bee colony; multi-objective optimization; external archive; opposition; elite learning.
Distance-based Immune Generalised Differential Evolution Algorithm for Dynamic Multi-Objective Optimisation
by María-Guadalupe Martínez-Peñaloza, Efrén Mezura-Montes, Alicia Morales-Reyes, Hernán Aguirre
Abstract: This paper presents Distance-based Immune Generalised Differential Evolution (DIGDE), an improved algorithmic approach to tackle dynamic multi-objective optimisation problems (DMOPs). Its novelty is using the Inverted Generational Distance (IGD) as an indicator in its selection mechanism to guide the search. DIGDE is based on the Immune Generalised Differential Evolution (Immune GDE3) algorithm which combines Differential Evolution (DE) fast convergence ability and Artificial Immune Systems (AIS) principles for good diversity preservation. A thorough empirical evaluation is carried out on novel benchmark problems configured with different dynamic characteristics. DIGDE's experimental results show an overall improved statistically supported performance in terms of solutions approximation and better achieved distributions. Using IGD as a searching indicator allows DIGDE to achieve better performance and robustness in comparison to state of-the-art methods when facing different change frequencies and severity levels.
Keywords: Dynamic Multi-objective Optimisation; Selection mechanism; Inverted Generational Distance indicator; Immune response; Differential Evolution; DMOPs.
A Novel Bio-Inspired Stochastic Framework to Solve Energy Management Problem in Hybrid AC-DC Microgrids with Uncertainty
by Sattar Shojaian
Abstract: This paper develops a new management framework for optimal operation of the hybrid AC-DC microgrids incorporating renewable energy sources and storages. Hybrid microgrid consists of two parts of AC and DC to supply the AC and DC loads, respectively. The power exchange capability of hybrid microgrids between the AC and DC parts makes it possible to reduce the total microgrid costs, effectively. To make it a realistic analysis, a stochastic method based on cloud theory is proposed to model the uncertainty effects of wind turbine power, photovoltaic power, load demand and market price sufficiently. The proposed framework makes use of a new optimization algorithm based on flower pollination mechanism to minimize the total network costs through the optimal dispatch of the units. Also, a three stage modification method is proposed to improve the population diversity and avoid the premature convergence. The performance of the proposed method is examined on the IEEE test system through two different operation scenarios.
Keywords: Evolutionary Algorithm; Energy Management; Hybrid Microgrid; Flower Pollination Algorithm; Cloud Theory; Modification; Uncertainty; Optimization.
Ensemble Learning Based Classification on Local Patches from Magnetic Resonance Images to Detect Iron Depositions in Brain
by Beshiba Wilson, Julia Punitha Malar Dhas, Ruma Madhu Sreedharan, Ram P. Krish
Abstract: Iron deposition in brain has been observed with normal aging and is associated with neurodegenerative diseases. The automated classification of brain Magnetic Resonance Images (MRI) based on iron deposition in Basal Ganglia region of brain has not been performed, to our knowledge. It is very difficult to analyze iron regions in brain using simple MRI techniques. The MRI sequence namely Susceptibility Weighted Imaging (SWI) helps to distinguish brain iron regions. The objective of our work is to investigate the iron regions in selected areas of basal ganglia region of brain and classify MR images. The study included a total of 60 MRI images which consists of 40 subjects with iron region and 20 subjects of healthy controls. We performed Gaussian smoothing followed by construction of 40 localized patches of each MR image based on iron and normal regions. Gray Level Co-occurrence Matrix (GLCM) features are extracted from the patches and fed to Random Forest (RF) Classifier for patch based classification of iron region. Training of data patch features was done by Random Forest Classifier and the performance of classifier in terms of accuracy was measured. The experimental results show that the proposed localized patch based approach for classification of brain iron images using Random Forest Classifier achieved 96.25% classification accuracy in identifying normal and iron regions from brain MR sequences.
Keywords: Ensemble Learning; Classification; Iron deposition; Magnetic Resonance Images (MRI); Susceptibility Weighted Images (SWI); Gray Level Co-occurrence Matrix (GLCM); Random Forest (RF) Classifier; Neurodegenerative diseases; Basal Ganglia; Gaussian Smoothing.
Fittest Survival: An Enhancement Mechanism for Monte Carlo Tree Search
by Jiajia Zhang, Xiaozhen Sun, Dandan Zhang, Xuan Wang, Shuhan Qi, Tao Qian
Abstract: Monte Carlo Tree Search (MCTS), which constructs a search tree of game states and evaluates expected rewards by thousands of Monte Carlo simulations, has become the pre-eminent approach for many challenging games. One severe challenge of MCTS is the contradiction between the accuracy of states' evaluation and practical time consumption for simulations, both of which are critical for a competitive game program. This paper proposes and evaluates Fittest Survival Monte Carlo Tree Search (FS-MCTS) which provides a novel mechanism to enhance MCTS towards correct direction with fewer simulations. The key idea of FS-MCTS is to keep states with significant advantages to survive while eliminate the others. This is the meaning of ``Fittest Survival''. In this sense, FS-MCTS no longer completely depends on the evaluation accuracy of game states which severely relies on adequate simulation times. We evaluate FS-MCTS in the problems of Poker. Experimental results show that FS-MCTS, combining with several variants of popular UCB policies, performs better than their vanilla versions when a certain number of simulations are guaranteed for its theoretical prerequisites.
Keywords: Monte Carlo Tree Search; Texas Poker; tree pruning; significance test.
Clustering algorithm for mixed attributes data based on Glowworm Swarm Optimization algorithm and K-prototypes algorithm
by Yaping Li, Zhiwei Ni, Weiliang Zhou
Abstract: The main purpose of this research is to improve the accuracy of clustering algorithm, which is also of great significance for big data analysis. As a typical clustering algorithm,K-prototypes algorithm directed at the analysis of mixed attributes data features simple principles and efficient operation. However, the clustering effect of K-prototypes algorithm depends on the initial clustering center, and the distance between data objects is not calculated accurately. In order to improve the clustering accuracy of mixed attributes data, Glowworm Swarm Optimization (GSO) algorithm is introduced into K-prototypes algorithm to form a new clustering algorithm in our research work. First, the idea of combining GSO algorithm and K-prototypes algorithm for the clustering of mixed attributes data is provided. Next, GSO algorithm is improved by using the good point set. Then, the improved GSO algorithm is employed to search extreme points of density in the space of data objects. The initial clustering center of K-prototypes algorithm is chosen from the extreme points of density. Meanwhile, a unified method is designed for the distance of numeric data and categorical data. On this basis, GSO algorithm and K-prototypes algorithm are combined to form a new clustering algorithm flow (GSOKP) for mixed attributes data. Finally, the UCI data sets of numeric data, categorical data and mixed data are selected to test GSOKP algorithm. And the effectiveness of GSOKP algorithm is analyzed in terms of clustering accuracy through experimental comparison.
Keywords: GSO Algorithm; K-prototypes Algorithm; Good Point Set; Clustering.
Adaptive Optimization Driven Deep Belief Networks for Lung Cancer Detection and Severity Level Classification
by Malayil Shanid, Anitha A
Abstract: Computed tomography (CT) for lung cancer detection is trending research in determining the lung cancer on its earlier stages. However, accurate Lung cancer detection with severity levels is a major challenge faced by most of the existing methods. This paper proposes a lung cancer detection model for analyzing the severity levels using the CT images. The input CT images are obtained from the input lung cancer database using which the lung cancer detection and severity level classification is performed. The shape local binary texture (SLBT) is employed, which is generated by combining Local directional pattern (LDP) and Linear Binary Pattern (LBP), which is extracted from the nodules. The features are subjected to proposed Adaptive-SEOA-DBN, which is the integration of Adaptive-Salp-Elephant Herding Optimization Algorithm (Adaptive-SEOA) in DBN for effective training of the model parameters. The proposed Adaptive-SEOA is developed by combining self-adaptive concept in the SEOA. Finally, severity level classification is done to declare the severity of patient. The effectiveness of the proposed Adaptive-SEOA-DBN is revealed based on maximal accuracy of 96.096 and minimal False Detection Rate (FDR) of 0.019, minimal False Positive Rate (FPR) of 4.999, and maximal True Positive Rate (TPR) of 96.096, respectively.
Keywords: Lung cancer; Severity level; CT images; segmentation; lung nodules.
Parameter Optimization of Sliding Window Algorithm Based on Ensemble Multi-objective Evolutionary Computation
by Guang Li, Jie Wang, Jing Liang, Caitong Yue, Taishan Lou
Abstract: The parameters of sliding window algorithm are difficult to determine. Therefore, a sliding window-based method for parameter optimization of data stream trend anomaly detection algorithm is proposed in this study. This method regards the data stream anomaly detection as a two-objective optimization problem. Three optimization algorithms and ensemble strategies were used to obtain the optimal parameter settings of the algorithm. With this strategy, it is no longer difficult to determine the parameters of the data stream trend anomaly detection algorithm based on the sliding window. Through verification of multiple real parameter data in Tarim Oilfield, it could be known that this method could realize the optimal parameter settings, which provides a reference for the parameter setting of the data stream trend anomaly detection algorithm based on sliding window.
Keywords: data stream; sliding window; parameter optimization; machine learning; anomaly detection; evolutionary computation; ensemble strategy; Multi-objective evolutionary; industry data; project parameters.
Special Issue on: Bio-inspired Evolutionary Computations and Their Applications
Brushless Direct Current Motor Design Using A Self-adaptive JAYA Optimization Algorithm
by Li Yan, Chuang Zhang, Boyang Qu, Kunjie Yu, Caitong Yue
Abstract: This paper proposes a self-adaptive JAYA (SAJAYA) to maximize the efficiency of the brushless direct current (BLDC) motor by optimizing the design parameters. In SAJAYA, an adaptive weight strategy is introduced into the original JAYA to control the degree of approaching the best solution and avoiding the worst solution at different evolution stages. Moreover, to maintain the diversity and enhance the local search ability, a new hybrid learning strategy is developed. Further, a self-adaptive selection mechanism is designed in order to dispatch the original learning strategy and the new hybrid learning strategy to each individual automatically. Based on this framework, the exploration and exploitation abilities of the population are expected to be balanced. Besides, a chaotic elite search is introduced to further refine the best solution of each generation. Experimental results show that the proposed SAJAYA shows a superior performance in solving the BLDC motor optimization problem compared with other well-established algorithms.
Keywords: Evolutionary computation; JAYA algorithm; Optimization; Brushless DC wheel motor.