Earticle

On the ultra-high speed broadband internet network transferring the massive multi-media data, the rapid restoration is very important. The online game industry is also fast growing with supporting of the real-time online network game. The online traffic on the internet network is steadily increasing. The optical based broadband convergence network provides these kinds of services. It is a form of multilayer network which loads the IP traffic on the lower optical DWDM. The failure of the link and node on this kind of network causes the significant loss of data. To solve this kind of problem, the new approach over the conventional one is necessary. It turned out to be its failure recovery thirty-five percent faster

We develop an intelligent information system in a multilayer electronic supply chain network. Using the internet for supply chain management (SCM) is a key interest for contemporary managers and researchers. It has been realized that the internet can facilitate SCM by making real time information available and enabling collaboration between trading partners. Here, we propose a multi-agent system to analyze the performance of the elements of a supply network based on the attributes of the information flow. Each layer consists of elements which are differentiated by their performance throughout the supply network. The proposed agents measure and record the performance flow of elements considering their web interactions for a dynamic route selection. A dynamic programming approach is applied to determine the optimal route for a customer in the end-user layer.

본 연구는 공공데이터포털을 통해 제공되는 2023년 건강검진정보 데이터를 머신러닝으로 학습하여 성별 생물학적 연령을 예측하는 판별 인공지능을 개발하는 것이다. 전체 데이터(100만명) 중 건강검진 결과가 측정된 330,752명(남자 179,556명, 여자 151,196명)을 머신러닝 학습 데이터로 활용했다. 연령대를 출력변수로, 키, 몸무게, 허리둘레, 수축기 혈압, 이완기 혈압, 공복 혈당, 중성지방, HDL 콜레스테롤, LDL 콜레스테롤, 헤모글로빈, 혈청 GOT, 혈청 GPT, 감마 GTP 등 26개 항목을 입력 변수로 사용하였다. 판별 인공지능 개발에는 머신러닝 알고리즘 중 딥러닝이 가능한 다층신경망을 사용하였으며, 인공지능의 성능평가에는 ROC 곡선을 사용하였다. 다층신경망으로 성별 생물학적 나이를 예측한 결과 입력변수(검강검진정보)를 학습하지 않은 학습데이터의 출력변수(연대기적 나이)의 연령분포와 입력변수를 학습한 인공지능이 예측한 출력변수(생물학적 나이)의 확률 값이 거의 비슷하게 나타났다. 인공지능 예측방법은 머신러닝 및 딥러닝을 이용하여 대용량 고차원 데이터에서 변수 간 비선형성과 복잡한 상호작용까지 포착하여 효율적으로 처리할 수 있다. 따라서, 다양한 임상 및 바이오마커 정보를 통합 분석함으로써 개인의 노화 정도를 정밀하게 파악할 수 있기 때문에, 예방 중심의 보건정책 수립에 중요한 과학적 근거를 제공할 수 있을 것이다.

This study aims to develop a discriminative artificial intelligence system that predicts gender and biological age by learning from 2023 health checkup information data provided through the public data portal using machine learning. Among the total data (1 million people), 330,752 individuals (179,556 men and 151,196 women) whose health screening results were measured were used as machine learning training data. Age group was used as the output variable, and 26 items, including height, weight, waist circumference, systolic blood pressure, diastolic blood pressure, fasting blood glucose, triglycerides, HDL cholesterol, LDL cholesterol, hemoglobin, serum GOT, serum GPT, and gamma GTP, were used as input variables. In the development of discriminatory artificial intelligence, a multilayer neural network capable of deep learning among machine learning algorithms was used, and the ROC curve was used to evaluate the performance of artificial intelligence. As a result of predicting biological age by gender using a multilayer neural network, the age distribution of the output variable (chronological age) of the training data that did not learn the input variable (health examination information) and the probability value of the output variable (biological age) predicted by the artificial intelligence that learned the input variable were almost similar. Artificial intelligence prediction methods can efficiently process large amounts of high-dimensional data by capturing nonlinearities and complex interactions between variables using machine learning and deep learning. Therefore, by integrating and analyzing various clinical and biomarker information, it is possible to precisely determine the degree of aging in an individual, thereby providing important scientific evidence for establishing prevention-oriented health policies.

미래전의 양상은 첨단과학기술의 발전에 따라 다양한 유ㆍ무인 전투체계들이 전투를 수행하는 것으로 변화하고 있다. 미래 의 무기체계들은 대량의 정보를 생성하고 송ㆍ수신하게 되며, 이에 따라 현재의 전술정보통신체계는 미래의 기반 통신체계로 운용하기에 제한된다. 또한 군단급 이상의 대규모 제대의 정보유통량을 분석하기 위해서는 전장에서 발생할 수 있는 다양한 상황들을 고려하여야 한다. 이에 본 논문에서는 Army TIGER 군단의 정보유통량을 분석하기 위해 지상망 단독 운용, 통합망 운용, 부하 분산, 장애 극복 시나리오 등 네 가지 시나리오를 기반으로 한 M&S 결과를 분석한다. 또한 지상ㆍ공중ㆍ우주망이 통합된 다계층 통합 전술통신망의 운용을 통해 네트워크의 신뢰성과 안정성을 향상할 수 있음을 제시한다.

Future warfare is evolving with advanced science and technology, introducing a variety of unmanned and manned com bat systems. These systems generate and exchange massive amounts of information, challenging current tactical communi cation systems as a foundation for future communication infrastructure. To analyze the information distribution capability of the Army TIGER corps, this paper examines four scenarios: standalone ground network operation, integrated network o peration, load distribution, and error recovery. Utilizing M&S results, we highlight the potential of a multilayer integrated command and control network, incorporating ground, air, and space networks, to enhance the reliability and stability of the overall communication network.

The performance of the aero-engine is an important safeguard to flying security. We can diagnose and predict the fault type through obtaining and analyzing the vibration signals based on the fault characters of the aero-engine. But, because of the complexity of the aero-engine’s structure, the vibration signals acquired from multiple groups of the piezoelectric sensors are often composed of several signal aliasing and other noise jamming, etc. Thus, the vibration signals are in the nonlinear or weak nonlinear state, the traditional blind source separation (BSS) algorithms usually adopt linear hypothesis to approximate equivalent to the nonlinear problems, which leading to the separation results not ideal even wrong. This paper applied a kind of multilayer perceptron BP neural network algorithm to realize the aero-engine vibration signal separation with high precision through the simulation and experiment, which proved this algorithm has a certain practical value to the aero-engine fault diagnosis and prediction.

This paper considers the problem of designing a Multilayer Survivable Optical Network for the customers’ Demands Problem called MSONDP. The network is modelled by two graphs: an undirected graph G1 = (V1, E1) and a complete undirected and weighted graph G2 = (V2, E2, c). The goal objective of this problem is to design connections based on customers’ demands with the smallest a minimum network cost to protect the network against all failures. This paper introduces a multi-objective approach for MSONDP. These objectives are to minimize the network cost (totalCost) and the maximum number of connections passing over a link (maxConn). Further, this paper also proposes a multi-objective genetic algorithm to solve this problem. The eExperimental results on real world and random instances are reported to show the efficiencyefficacy, in terms of minimizing the network cost, of the proposed algorithm comparing compared to the single genetic PGAMSONDP.

In order to improve universality of detection effect of complex network covert channel, coarsening clustering of channel is achieved, based on clustering algorithm, with hierarchical clustering in this article, then recognition detection of two-stage channel refinement is carried out in coarsening clustering results of each layer based on density clustering. Clustering method of multilayer covert channel is designed based on density, the effectiveness of proposed method is verified by carrying out contrast experiment on density clustering method, entropy method and  evaluation method; in the end, experiments of simulation example demonstrate that such algorithm is able to quickly and accurately detect covert channel of complex network when the noise is not higher than 20%.

In this paper, a pre-layout design approach for high-speed microprocessor is proposed. For multilayer PCB stark up configuration as well as selection and placement of decoupling capacitors, an effective solution for reducing SSN and EMI is obtained by modeling and simulation of complete power distribution system. The system model includes VRM, decoupling capacitors, multiple power and ground planes for core voltage, vias, as well as microprocessor. Finally, the simulation results are verified by measurements data.

Traditional deterministic channel modeling is accurate in prediction, but due to its complexity, improving computational efficiency remains a challenge. In an alternative approach, we investigated a multilayer artificial neural network (ANN) to predict large-scale and small-scale channel characteristics in metro tunnels. Simulated high-precision training datasets were obtained by combining measurement campaign with a ray tracing (RT) method in a metro tunnel. Performance on the training data was used to determine the number of hidden layers and neurons of the multilayer ANN. The proposed multilayer ANN performed efficiently (10 s for training; 0.19 ms for prediction), and accurately, with better approximation of the RT data than the single-layer ANN. The root mean square errors (RMSE) of path loss (2.82 dB), root mean square delay spread (0.61 ns), azimuth angle spread (3.06°), and elevation angle spread (1.22°) were impressive. These results demonstrate the superior computing efficiency and model complexity of ANNs.

We develop an intelligent information system in a multilayer electronic supply chain network. Using the internet for supply chain management (SCM) is a key interest for contemporary managers and researchers. It has been realized that the internet can facilitate SCM by making real time information available and enabling collaboration between trading partners. Here, we propose a multi-agent system to analyze the performance of the elements of a supply network based on the attributes of the information flow. Each layer consists of elements which are differentiated by their performance throughout the supply network. The proposed agents measure and record the performance flow of elements considering their web interactions for a dynamic route selection. A dynamic programming approach is applied to determine the optimal route for a customer in the end-user layer.

This study analyzed the combinations of adverbs and predicates(verbs and adjectives) in Korean using the ‘Lexical Semantic Analysis Corpus(어휘의미분석 코퍼스)’ provided by the National Institute of Korean Language(국립국어원). By distinguishing between written and spoken language, the study identified frequent combinations such as ‘안-되다’. Additionally, Cramer's V coefficient was employed to measure the semantic relationships between adverbs and verbs, and a multi-layer network analysis was conducted to categorize Korean verbs into 17 clusters. This research offers a novel perspective on Korean verb studies and is expected to contribute to future research in natural language processing.

동적 신경망은 temporal 신호처리가 요구되는 여러 분야에 사용되어 왔다. 본 논문에서는 다층 신경망의 동특성을 향상시키기 위해 감마 신경망(GAM) 다루고 있다. GAM 신경망은 순방향 다층 신경망의 히든층에 감마 메모리 커널을 사용하고 있다. GAM 신경망은 선형 및 비선형 시스템 식별을 통해 평가되었으며 상대적인 성능평가를 위해 순방향 신경망(FNN)과 리커런트 신경망(RNN)과 비교하고 있다. 실험결과에 의하면 GAM 신경망은 학습속도와 정확도에서 더 우수하게 동작하였으며, 이러한 사실은 시스템 식별에 있어서 GAM 신경망이 기존의 다른 다층 신경망보다 더 효과적인 신경망이 될 수 있음을 보여주었다.

Dynamic neural networks have been applied to diverse fields requiring temporal signal processing. This paper presents gamma neural network(GAM) to improve the dynamics of multilayer network. The GAM network uses the gamma memory kernel in the hidden layer of feedforword multilayer network. The GAM network is evaluated in linear and nonlinear system identification, and compared with feedforword(FNN) and recurrent neural networks(RNN) for the relative comparison of its performance. Experimental results show that the GAM network performs better with respect to the convergence and accuracy, indicating that it can be a more effective network than conventional multilayer networks in system identification.

동적 신경망은 temporal 신호처리가 요구되는 여러 분야에 사용되어 왔다. 본 논문에서는 다층 신경망의 동특성을 향상시키기 위해 감마 신경망(GAM) 다루고 있다. GAM 신경망은 순방향 다층 신경망의 히든층에 감마 메모리 커널을 사용하고 있다. GAM 신경망은 선형 및 비선형 시스템 식별을 통해 평가되었으며 상대적인 성능평가를 위해 순방향 신경망(FNN)과 리커런트 신경망(RNN)과 비교하고 있다. 실험결과에 의하면 GAM 신경망은 학습속도와 정확도에서 더 우수하게 동작하였으며, 이러한 사실은 시스템 식별에 있어서 GAM 신경망이 기존의 다른 다층 신경망보다 더 효과적인 신경망이 될 수 있음을 보여주었다.

Dynamic neural networks have been applied to diverse fields requiring temporal signal processing. This paper presents gamma neural network(GAM) to improve the dynamics of multilayer network. The GAM network uses the gamma memory kernel in the hidden layer of feedforword multilayer network. The GAM network is evaluated in linear and nonlinear system identification, and compared with feedforword(FNN) and recurrent neural networks(RNN) for the relative comparison of its performance. Experimental results show that the GAM network performs better with respect to the convergence and accuracy, indicating that it can be a more effective network than conventional multilayer networks in system identification.

대규모 병렬처리가 가능하고 칩당 뉴론 집적도가 높은 펄스형 디지털 다계층 신경망 구조를 제안하였다. 제안된 신경망에서는 대수적인 신경망연산이 의사-랜덤 펄스 시퀀스(pseudo-random pulse sequences)와 단순 디지털 논리 게이트를 이용하여 확률적 프로세스로 대치되었다. 확률적 프로세스의 결과로 나타나는 신경망 연산의 통계적 모델을 제시하였으며 이를 바탕으로 랜덤잡음의 영향과 연산의 정확도를 분석하였다. 이진인식 문제를 적용하여 제안된 신경망의 성능을 평가하고 제시한 통계적 분석결과의 정당성을 검증하였다. Gate 레벨과 register transfer 레벨로 기술된 신경망의 VHDL 모델의 시뮬레이션 결과는 개발된 통계적모델로 예측된 인식추정치와 실제 인식률이 거의 일치함을 보였으며, 또한 숫자인식률에 있어서도 일반 Back-Propagation 신경망의 인식률과 거의 차이가 없음을 보였다.

In this paper, a pulse-mode digital multilayer neural network with a massively parallel yet compact and flexible network architecture is presented. Algebraicneural operations are replaced by stochastic processes using pseudo-random pulse sequences and simple logic gates are used as basic computing elements. The distributions of the results from the stochastic processes are approximated using the hypergeometric distribution. A statistical model of the noise(error) is developed to estimate the relative accuracy associated with stochastic computing in terms of mean and variance. Numerical character recognition problems are applied to the network to evaluate the network performance and to justify the validity of analytic results based on the developed statistical model. The network architectures are modeled in VHDL using the mixed descriptions of gate-level and register transfer level (RTL). Experiments show that the statistical model successfully predicts the accuracy of the operations performed in the network and that the character classification rate of the network is competitive to that of ordinary Back-Propagation networks.

This paper proposes an efficient method for improving the training performance of the neural network by using a hybrid of a stochastic approximation and a backpropagation algorithm. The proposed method improves the performance of the training by appliying a global optimization method which is a hybrid of a stochastic approximation and a backpropagation algorithm. The approximate initial point for a stochastic approximation and a backpropagation algorihtm. The approximate initial point for fast global optimization is estimated first by applying the stochastic approximation, and then the backpropagation algorithm, which is the fast gradient descent method, is applied for a high speed global optimization. And further speed-up of training is made possible by adjusting the training parameters of each of the output and the hidden layer adaptively to the standard deviation of the neuron output of each layer. The proposed method has been applied to the parity checking and the pattern classification, and the simulation results show that the performance of the proposed method is superior to that of the backpropagation, the Baba's MROM, and the Sun's method with randomized initial point settings. The results of adaptive adjusting of the training parameters show that the proposed method further improves the convergence speed about 20% in training.

In this paper, we describe the algorithm which controls an unknown nonlinear system with disturbance a using multilayer neural network. The multilayer neural network can be used to approximate any continuous function to any desired degree of accuracy. With the former fact, we approximate an unknown nonlinear system by using of multilayer neural netowrk. WE include a disturbance among the modelling error, and the weight-update rule of multilayer neural network is derived to satisfy Laypunov stability. The whole control system constitutes controller using the feedback linearization method. The weight of neural network which is used to implement nonlinear function is updated by the derived update-rule. The proposed control algorithm is verified through computer simulation.

This paper was written to explain Saroguk's foreign relations network through the distribution of ironware. By examining the ironware consumed in tombs that have spread across the eastern part of the Nakdong River, the Yeongnam region, and the Midwestern region from the Wooden Compartment Tomb Stage, the Saroguk-led ironware distribution network was diagnosed, and the differences in managerial patterns varied by JinHan(辰韓), Jin･ByeonHan(辰･弁韓), and MaHan(馬韓) regions were pointed out. The cast iron axe with triple~fourth protruding bands(3~4條凸帶鑄造鐵斧), iron tool with a swirly ornament(有刺利器), and the Ceremonial Iron Spearheads, which spread in earnest from the Wooden Compartment Tomb Stage, are recognized as items that share and imitate a specific form expressed in Saroguk. In particular, the iron tool with a swirly ornament and iron spearhead for burial (branched bars type, type with a swirly ornament, and plate-shaped type) are noted for establishing the unique Ceremonial style of Saroguk. Even within the distribution range, the branched bars-Iron Spearhead and plate-shaped Iron Spearhead are spread over a wide area in the eastern part of the Nakdong River and the Midwest region, excluding the western area of the Nakdonggang River, while the iron tool, iron spearhead with a swirly ornament are distributed only in the eastern part of the Nakdong River with the Saro area as the core. As such, it is important that even if the style of the Saroguk's Ceremonial Ironware is collected, it is organically combined with the spread of store rituals(Deployment in 屍臺, the head of a department, etc.) while the distribution range varies according to the symbolism of each purpose or use of the Saroguk.

많은 제어기법들이 이산시간영역제어계에서 제어성능을 개선하기 위해서 제안되고 있다. 이 제어기법들을 이용한 제어계에서 계의 응답특성은 제어기의 이득에 관계한다. 특히 외란이나 부하변동에 의해서 계의 응답이 변할 때 제어기의 이득을 재조정할 필요가 있다. 본 논문에서는 다층 신경망으로 I-PD제어기와 전치보상기를 설계하였다. I-PD제어기와 전치보상기의 이득이 자동적으로 역전파 알고리즘에 의해서 조정되도록 하였다. 제어계의 응답이 어떤 조건에 의해서 변할 때 I-PD제어기와 전치보상기의 이득들이 역전파 알고리즘에 의해서 자동적으로 조정되게 하였다. 이 I-PD제어기법을 직류 서보 전동기를 구동원으로 하는 위치제어계에 적용하여 제어기의 제어성능을 실험 결과로 타당성을 확인하였다.

Many control techniques have been proposed in order to improve the control performance in discrete-time domain control system. In control system using these techniques, the response-characteristic of system is dependent on the gains of the controller. Specially, There is a need to readjust the gain of controller when the response of system is changed by disturbance or load fluctuation. In this paper, I-PD controller and pre-compensator are designed by multilayer neural network. The gains of I-PD controller and pre-compensator are adjusted automatically by back propagation algorithm when the response characteristic of system is changed under a condition. Applying this control technique to the position control system using a DC servo motor as a driver, the control performance of controller is verified by the results of experiment.

As an approach to develope a control system with robustness in changing control environment conditions, this paper will propose a robot manipulator control system using multilayer neural network and immune algorithm. The proposed immune algorithm which has the characteristics of immune system such as distributed and anomaly detection, probabilistic detection, learning and memory, consists of the innate immune algorithm and the adaptive immune algorithm. We will demonstrate the effectiveness of the proposed control system with simulations of a 2-link robot manipulator.

The use of system identification approaches for damage detection has been expanded in recent years. Soft computing techniques such as neural networks have been utilized increasingly. Damage assessment using neural networks is presented in this study. Data set for training neural networks are acceleration response of simple beam under the various damage states ,which are the inputs. The outputs are the damage locations and extents. Not only the trained damages but also untrained damages are. detected accuratelyintheassessmentstage.