Earticle

Medical image processing is one of the most challenging and emerging filed. Processing of medical image is one of the important tasks for the diagnosis of brain tumor. Image segmentation is required for detection of brain tumors, which is a quite complicated job if performed automatically. In recent time, scientists from various fields including medical, mathematical and computer science have collaborated to find out a better understanding of the disease and devise more cost-effective treatments. Due to advancements in the field of science and technology, we have innumerous methods for image segmentation which are used for the detection of brain tumor and to clearly recognize it from MRI imagery. Various methods and algorithms have been implemented for segmenting MRI imagery. This work implements particle swarm optimization technique to recognize brain tumor by characterizing MRI images. Machine learning algorithm is used for severity analysis of brain tumor.

Haptic actuators for large display panels play an important role in bridging the gap between the digital and physical world by generating interactive feedback for users. However, the generation of meaningful haptic feedback is challenging for large display panels. There are dead zones with low haptic sensations when a small number of actuators are applied. In contrast, it is important to control the traveling wave generated by the actuators in the presence of multiple actuators. In this study, we propose a particle swarm optimization (PSO)-based algorithm for the haptic localization of plates with electrostatic vibration actuators. We modeled the transverse displacement of a plate under the effect of actuators by employing the Kirchhoff-Love plate theory. In addition, starting with twenty randomly generated particles containing the actuator parameters, we searched for the optimal actuator parameters using a stochastic process to yield localization. The capability of the proposed PSO algorithm is reported and the transverse displacement has a high magnitude only in the targeted region.

Orthogonal Frequency Division Multiplexing (OFDM) is the best choice for wireless and wired high data rate communications. The OFDM system has many advantages such as high spectral efficiency, robustness to channel fading, and immunity to impulse interference. However, it has a main drawback, which is the Peak to Average Power Ratio (PAPR). Selecting Mapping (SLM) and Partial Transmit Sequences (PTS) are the two essential techniques for reducing the PAPR; however, they need to send side information to indicate how the transmitter generates signals. In this paper, Particle Swarm Optimization (PSO)-based dummy sub-carriers are inserted with the data to reduce PAPR without transmitting side information. The incorporation of PSO-based Dummy Sub-carriers Insertion (DSI) method on this system is performed by measuring the PAPR at the IFFT output (Inverse Fast Fourier Transform) after adding 6 adaptive sub-carrier sequences to the input data of IFFT. The PSO-based dummy algorithm decides to send the data once the PAPR of this data sequence and dummy sub-carrier is under a specific threshold. Using MATLAB as a mathematical simulation tool, simulation results are presented for a standard OFDM network model. The network model has been simulated in an Additive White Gaussian Noise (AWGN) channel environment to investigate PAPR and Bit-Error Rate (BER) performance. The simulation results show that the dummy sequence based on PSO reduces PAPR down to 4 dB compared to a conventional OFDM system and down to 3 dB compared with the clipping technique for the same BER performance.

Task scheduling is one of the most important issues in heterogeneous environments when high efficiency is required. Because task scheduling is a Nondeterministic Polynomial (NP)-hard problem, many evolutionary algorithms have been adopted to solve this problem. Since the convergence speed of solutions in population-based algorithms is low, they are integrated with local search algorithms. Thus, in this paper, to optimize the task scheduling makespan, a hybrid particle swarm optimization and hill climbing algorithm is proposed. The experimental results on random and scientific Directed Acyclic Graph (DAG) showed that the proposed algorithm performs effectively in terms of the makespan compared to the current well-known heuristic and particle swarm optimization algorithms.

The dorsal hand vein biometric system developed has a main objective and specific targets; to get an electronic signature using a secure signature device. In this paper, we present our signature device with its different aims; respectively: The extraction of the dorsal veins from the images that were acquired through an infrared device. For each identification, we need the representation of the veins in the form of shape descriptors, which are invariant to translation, rotation and scaling; this extracted descriptor vector is the input of the matching step. The optimization decision system settings match the choice of threshold that allows accepting/rejecting a person, and selection of the most relevant descriptors, to minimize both FAR and FRR errors. The final decision for identification based descriptors selected by the PSO hybrid binary give a FAR =0% and FRR=0% as results.

Training neural networks is a complex task with great importance in the field of supervised learning. In the training process, a set of input-output patterns is repeated to an artificial neural network (ANN). From those patterns weights of all the interconnections between neurons are adjusted until the specified input yields the desired output. In this paper, a new hybrid algorithm is proposed for global optimization of connection weights in an ANN. Dynamic swarms are shown to converge rapidly during the initial stages of a global search, but around the global optimum, the search process becomes very slow. In contrast, the gradient descent method can achieve faster convergence speed around the global optimum, and at the same time, the convergence accuracy can be relatively high. Therefore, the proposed hybrid algorithm combines the dynamic particle swarm optimization (DPSO) algorithm with the backpropagation (BP) algorithm, also referred to as the DPSO-BP algorithm, to train the weights of an ANN. In this paper, we intend to show the superiority (time performance and quality of solution) of the proposed hybrid algorithm (DPSO-BP) over other more standard algorithms in neural network training. The algorithms are compared using two different datasets, and the results are simulated.

The different issues in the wireless channel, such as multipath fading, shadowing, and hidden terminal problems, limit the sensing abilities of the single user in a cognitive radio network (CRN). In this connection, it is more suitable to follow cooperative spectrum sensing (CSS). However, the various categories of malicious users (MUs) pose threats to the performance of CSS. These MUs are eager to report false sensing data to the fusion center (FC) to deceive its decision. This work considered the CSS performance in the presence of always yes MU (AYMU), opposite MU (OMU), and always no MU (ANMU). Moreover, the considered particle swarm optimization (PSO) algorithm takes out reports of the MUs using one to many sensing distances (OTMSD) and z-score outlier detection techniques before the final decision. Simulation results validate the effectiveness of the proposed scheme in the presence of MUs.

Machine learning-based sentiment classification is the best-performing method to understand public sentiment. However, the method has some problems, such as noisy features and high-dimensional feature space which affect the sentiment classification performance. To address the problems, this paper proposes a new feature selection using hybrid Query Expansion Ranking and Binary Particle Swarm Optimization with Adaptive Inertia Weights. The proposed method was validated using five tweet datasets on different topics both in Indonesian and English, and compared with state-of-the-art of filter and wrapper-based feature selection methods. Experimental results show the proposed method significantly improves sentiment classification performance and decrease computational time.

Orthogonal frequency division multiplexing (OFDM) is a multicarrier modulation technique for high-speed data transmission in wireless communications. In an OFDM system, a large number of subcarriers are used to transmit the modulated symbols and consequently, the OFDM signals have a high peak-to-average power ratio (PAPR). To reduce the high PAPR, we propose a partial transmit sequence (PTS) method based on the adaptive particle swarm optimization. Also, the proposed method efficiently search the optimal combination of phase rotation factors to decrease the computational complexity. Experimental results show that the proposed method significantly has reduced the PAPR and computational complexity.

건설사업 초기단계에서 제한된 사업정보를 가지고 개략적인 공사비를 예측하는 것은 사업의 수익성과 다른 사업 추진을 위한 자금의 여유 등을 고려할 때 발주처나 사업주체 같은 건설사업 관계자에게 중요하다. 이에 통계적인 방법과 인공지능기법을 이용하여 사업 초기단계에서의 공사비 예측에 관한 많은 연구가 이루어졌다. 본 논문은 새나 물고기 무리의 군집행동에 영감을 얻은 최적의 해를 찾아가는 진화형 계산기법인 PSO방법을 이용하여 공동주택 건설사업 초기단계에서의 공사비를 예측하는 새로운 방법을 제시한다. NeuroShell Predictor 소프트웨어를 이용하여 219개 공동주택 건설사업자료를 가지고 예측모델을 제안하였다. 소프트웨어에서 제공하는 성능에 대한 통계치와 기존의 연구방법으로 많이 적용된 신경망모델과의 직접적인 비교를 통해서 제안된 PSO기반 예측모델의 성능은 좋은 것으로 판명되었다. 또한, 본 연구에서의 성공적인 적용을 통해서 다른 예측문제의 해결에 있어서 PSO방법의 잠재능력을 보여주었다.

At the initial stage of construction projects, predicting approximate cost with limited information about projects is important to project participants such as owners and developers in that they can estimate expected profits from projects and available funds for other organizational activities. From this point, many researches to predict construction costs at the initial stage have been performed using statistical methods and artificial intelligence methods. This paper proposes a new approach to predict construction costs at the initial stage of apartment housing projects using a PSO(Particle Swarm Optimization) method, a population-based stochastic optimization one inspired by social behaviors of bird flocking or fish schooling. Using the NeuroShell Predictor software, the PSO-based prediction model was suggested with 219 apartment housing projects whose data were obtained from bidding announcements for construction supervisors in the Korea Construction Consulting Engineers Association. From high R-squared and correlation coefficient values provided by the software and in comparison with an artificial neural network(ANN) model, the performance of the suggested prediction model was found to be good. And the results showed that the PSO method can have the potential to provide valuable solutions to other prediction problems.

본 논문에서는 휴먼 행동 인식에 있어 수집 데이터의 특성에 따라 기존 PSO (Particle Swarm Optimization) 알고리즘의 경계선 부분에서 발생되는 문제을 군집들 사이의 거리를 기반으로 개선하는 알고리즘을 제안한다. PSO를 이용한 휴먼 행동 인식 방법은 입력 데이터를 경계값과 비교하여 인식하므로 상으로 빠르고 간단한 방 법이다. 하지만, 착용형 장치와 센서를 이용하는 휴먼 행동 인식에서와 같이 데이터 수집 환경의 변화에 따라 데이 터 특성의 변화가 심하여 학습 데이터와 테스트 데이터 사이의 차이가 생기는 경우 경계값 내에 포함되지 못하는 입 력 데이터들로 인해 성능이 하된다. 이를 해결하기 해 본 논문에서는 경계를 벗어나는 입력 데이터에 해 입력 데이터와 각 군집들의 경계와의 거리를 이용하여 행동 인식 성능을 개선한다. 실험 결과 앉기, 서기, 그리고 걷기와 같은 3 가지 휴먼 행동에 해 SVM (Support Vector Machine) MLP (Multi-Layer Perceptron)과 기존 PSO와 비교하여 최 약 18%의 인식 성능 개선을 가져오는 것을 확인하다.

This paper proposes a modified PSO (Particle Swarm Optimization) algorithm for human activity recognition, which employs distance between boundaries to solve a problem arisen from the boundaries of the existing PSO algorithm. This problem generally caused by inhomogeneous characteristics of the collected human activity data in daily life. Human activity recognition algorithm based on PSO is simple and fast since it classifies input data by comparing boundaries only. However, changes of data collection environment for human activity recognition using wearable and sensor devices cause changes in characteristics of input data, and this results in increase of input data outside of boundaries and decrease of classification performance according to different characteristics between training data and test data. In order to solve this problem, this paper increases performance by correcting boundaries of clusters using the distance between input data and every clusters for input data outside of every boundaries. Experimental results show that the proposed method demonstrates maximum 18% better performance comparing with SVM (Support Vector Machine), MLP (Muti-Layer Perceptron) and the existing PSO respectively for 3 human activities such as sitting, standing, and walking.

근래에 게임이론 분야에서 진화계산법을 사용한 교섭게임 분석은 중요한 이슈 중에 하나이다. 본 논문에서는 이질적인 두 인공 에이전트 간의 공진화를 활용하여 교섭게임을 관찰한다. 두 인공 에이전트를 모델링하기 위해 사용된 전략은 진화전략의 종류인 입자군집최적화와 차분진화알고리즘이다. 교섭게임에서 각 전략이 최선의 결과를 얻기 위한 알고리즘 모수들을 조사하고 두 전략의 공진화를 관찰하여 어느 알고리즘이 교섭게임에 더 우수한지 관찰한다. 컴퓨터 시뮬레이션 실험 결과 입자군집최적화 전략이 차분진화알고리즘 전략보다 교섭게임에서 더 우수한 성능을 보임을 확인하였다.

Recently, analysis of bargaining game using evolutionary computation is essential issues in field of game theory. In this paper, we observe a bargaining game using co-evolution between two heterogenous artificial agents. In oder to model two artificial agents, we use a particle swarm optimization and a differential evolution. We investigate algorithm parameters for the best performance and observe that which strategy is better in the bargaining game under the co-evolution between two heterogenous artificial agents. Experimental simulation results show that particle swarm optimization outperforms differential evolution in the bargaining game.

죄수 딜레마 게임은 게임 이론의 대표적인 사례로 많은 경제학자, 사회과학자 및 컴퓨터 과학자가 관심을 가지고 연구하고 있다. 근래에는 죄수 딜레마 게임 분석을 위해 유전 알고리즘, 입자 군집 최적화 등의 진화 연산 기법을 적용한 계산적 접근에 대한 연구가 활발히 이루어져 왔다. 본 연구에서는 3가지의 서로 다른 이진입자 군집 최적화 기법을 사용하여 2명 또는 그 이상의 플레이어가 참여하는 반복 죄수 딜레마 게임에 대한 전략을 진화시켜보고자 한다. 반복 죄수 딜레마 게임에 3가지 버전의 이진 입자 군집 최적화를 적용하여 실험한 결과 자신의 이득을 최대화하기 위한 이기적인 참가들 사이에서도 상호 협력 관계가 구축될 수 있음을 확인하였나 참여자가 많을수록 상호 협력 관계가 구축이 어려워 짐을 확인하였다.

The prisoner's dilemma game which is a representative example of game theory is being studied with interest by many economists, social scientists, and computer scientists. In recent years, many researches on computational approaches that apply evolutionary computation techniques such as genetic algorithms and particle swarm optimization have been actively conducted to analyze prisoner dilemma games. In this study, we intend to evolve a strategy for a iterated prisoner dilemma game participating two or more players using three different binary particle swarm optimization techniques. As a result of experimenting by applying three kinds of binary particle swarm optimization to the iterated prisoner's dilemma game, it was confirmed that mutual cooperation can be established even among selfish participants to maximize their own gains. However, it was also confirmed that the more participants, the more difficult to establish a mutual cooperation relationship.

In this paper, the particle swarm optimization (PSO) algorithm is adopted to design a modified ring-slot type patch rectenna with a resonance frequency of 2.45GHz. In order to accomplish minimization of dimensions and circular polarization (CP) and harmonic suppression, axial direction slits and side-cuts are added to the patch of the ring. The PSO manipulated this kind of multi-dimensional problem very well, and as a result, the designed rectenna shows a desirable performance of return loss of 21.36dB and axial ratio of 2.92dB at the frequency of 2.45GHz with compact sizing.

본 논문에서는 기존 파티클 스웜 최적화를 기반으로 추적 대상 객체의 이동 궤적을 이용하는 객체 추적기에서 시간 정보 활용의 문제점을 개선한 강인한 객체 추적 알고리즘을 제안한다. 제안하는 알고리즘은 추적 대상 객체와 유사 한 특징을 가지는 변위들의 집합에 대한 위치들의 온라인 업데이트와 추적을 가능하게 한다. 객체들의 중첩을 검출 하고 추적 대상의 위치를 결정하기 위해 궤적 정보와 변위들의 집합을 기반으로 적응적 파라미터를 사용하는 규칙 기반 접근을 사용한다. 기존 알고리즘들과 비교해보면 제안하는 접근법은 가용한 정보를 복합적으로 사용함으로써 각종 임계값에 대한 적응적 조정을 가능하게 한다. 또한, 파티클 스웜 최적화에서 발산에 의한 손실과 불완전한 수 렴의 문제를 해결하기 위해 효율적인 가중치 조절 함수를 제안하고 있다. 제안하는 가중치 조절 함수는 파티클들이 최적의 해에 수렴하기 이전에 전체 프레임 영역에서 탐색할 수 있도록 한다. 유사한 특징 조합을 가지는 다중 객체 가 존재하는 환경에서 제안 알고리즘을 테스트한 결과, 기존 스웜 최적화 기반의 객체 추적기들에 비해 기존 유사 변위들에 대한 잘못된 추적을 현저히 줄이는 것을 확인할 수 있었다.

This paper proposes an enhanced object tracking algorithm to compensate the lack of temporal information in existing particle swarm optimization based object trackers using the trajectory of the target object. The proposed scheme also enables the tracking and documentation of the location of an online updated set of distractions. Based on the trajectories information and the distraction set, a rule based approach with adaptive parameters is utilized for occlusion detection and determination of the target position. Compare to existing algorithms, the proposed approach provides more comprehensive use of available information and does not require manual adjustment of threshold values. Moreover, an effective weight adjustment function is proposed to alleviate the diversity loss and pre-mature convergence problem in particle swarm optimization. The proposed weight function ensures particles to search thoroughly in the frame before convergence to an optimum solution. In the existence of multiple objects with similar feature composition, this algorithm is tested to significantly reduce convergence to nearby distractions compared to the other existing swarm intelligence based object trackers.

본 논문에서는 커널 Extreme Learning Machine을 기반으로 하여 최적화 기법들 중의 하나인 입자 군집 최적화 기법을 이용한 설계 기법을 제시한다. 제안된 Kernel Extreme Learning Machine은 기존에 사용되어지는 뉴럴 네트워크의 단점을 개선한 네트워크이다. 다시 말하면, 뉴럴 네트워크의 단점인 느린 학습속도를 개선한 네 트워크이다. 일반적으로 뉴럴 네트워크의 히든 노드들은 랜덤 초기화 후 오류 역전파 알고리즘을 이용하여 학습 한다. 이와 같은 오류 역전파 알고리즘은 매우 느린 학습속도를 보인다. 이와 같은 단점을 해결하기 위하여, Kernel Extreme Learning Machine의 히든 노드들은 랜덤 초기화 되고 학습되지 않고 출력층의 연결 하중만 학습 되어진다. 이와 같은 장점을 가진 Kernel Extreme Learning Machine의 구조를 최적화하기 위하여 입자 군집 최적 화 기법을 사용한다. 제안된 설계 방법을 적용하여 설계된 모델의 일반화 성능의 우수성을 보이기 위하여, 다수 의 머신러닝 데이터들을 이용하여 실험하고 실험을 통해 얻은 결과를 비교 평가하였다.

In this paper, we proposed the design method of Extreme Learning Machine which is optimized by using Particle Swarm Optimization Technique. Extreme Learning Machine is the improved version of the conventional neural networks which have a very slow learning speed based on the back-propagation algorithm. In the conventional neural networks, the connection weights between the input layer and the hidden layers are initialized randomly and then optimized by using the gradient decent method. The speed of the learning method for the conventional neural networks is slow. In order to overcome the drawback of the conventional neural networks, the connection weights of the hidden nodes are just initialized randomly and will not be optimized, and the only connection weights of the output nodes are learned by using least square estimation not the iterative learning method. In addition, we use Particle Swarm Optimization to optimize the proposed Extreme Learning Machine. Several machine learning bench-mark data sets are used to show the generalization performance of the proposed design method and to compare their performance with the other already studied models.

In this paper, a study was conducted on the methodology for learning the parameters of a neural network using an evolutionary algorithm such as a particle swarm optimization algorithm. The possibility of using the particle swarm optimization algorithm for deep learning was analyzed, and various methods were considered for practical use.