• 전자통신동향분석
• /
• 제35권4호
• /
• pp.115-134
• /
• 2020

Several studies on machine learning (ML) based self-organizing networks (SONs) have been conducted, specifically for LTE, since studies to apply ML to optimize mobile communication systems started with 2G. However, they are still in the infancy stage. Owing to the complicated KPIs and stringent user requirements of 5G, it is necessary to design the 5G SON engine with intelligence to enable users to seamlessly and unlimitedly achieve connectivity regardless of the state of the mobile communication network. Therefore, in this study, we analyze and summarize the current state of machine learning studies applied to SONs as solutions to the complicated optimization problems that are caused by the unpredictable context of mobile communication scenarios.

• 한국생산제조학회지
• /
• 제20권3호
• /
• pp.310-315
• /
• 2011

General portal machine represents a distinct weak spot concerning their structural behavior because of long protruding structure components, such as saddles and rams. The weak point causes the deformation of the machine tool and consequently rises a severe machining error. The purpose of this study is to improve the structural design of crossrail in order to minimize it's distortion. Tool Center Point (TCP) was chosen as a reference point for evaluating the distortion effect of a crossrail and topological optimization was adopted as a method of structural design improvement. The displacements of TCP according to the machining positions were investigated by structural analyses for both of original crossrail design and the improved one. The comparing results showed that the displacement of TCP could be reduced about 55% maximum.

• Journal of Electrical Engineering and Technology
• /
• 제9권2호
• /
• pp.379-389
• /
• 2014

Low frequency oscillations (LFOs) are load angle oscillations that have a frequency between 0.1-2.0 Hz. Power system stabilizers (PSSs) are very effective controllers in improvement of the damping of LFOs. PSSs are designed by linearized models of the power system. This paper presents a new model of the power system that has the advantages of the Single Machine Infinite Bus (SMIB) system and the multi machine power system. This model is named a single machine normal-bus (SMNB). The equations that describe the proposed model have been linearized and a lead PSS has been designed. Then, particle swarm optimization technique (PSO) is employed to search for optimum PSS parameters. To analysis performance of PSS that has been designed based on the proposed model, a few tests have been implemented. The results show that designed PSS has an excellent capability in enhancing extremely the dynamic stability of power systems and also maintain coordination between PSSs.

• 전자통신동향분석
• /
• 제34권4호
• /
• pp.32-42
• /
• 2019

The performance of machine learning algorithms significantly depends on how a configuration of hyperparameters is identified and how a neural network architecture is designed. However, this requires expert knowledge of relevant task domains and a prohibitive computation time. To optimize these two processes using minimal effort, many studies have investigated automated machine learning in recent years. This paper reviews the conventional random, grid, and Bayesian methods for hyperparameter optimization (HPO) and addresses its recent approaches, which speeds up the identification of the best set of hyperparameters. We further investigate existing neural architecture search (NAS) techniques based on evolutionary algorithms, reinforcement learning, and gradient derivatives and analyze their theoretical characteristics and performance results. Moreover, future research directions and challenges in HPO and NAS are described.

• ETRI Journal
• /
• 제45권4호
• /
• pp.594-602
• /
• 2023

Internet of things (IoT) is commonly employed to detect different kinds of diseases in the health sector. Systemic lupus erythematosus (SLE) is an autoimmune illness that occurs when the body's immune system attacks its own connective tissues and organs. Because of the complicated interconnections between illness trigger exposure levels across time, humans have trouble predicting SLE symptom severity levels. An effective automated machine learning model that intakes IoT data was created to forecast SLE symptoms to solve this issue. IoT has several advantages in the healthcare industry, including interoperability, information exchange, machine-to-machine networking, and data transmission. An SLE symptom-predicting machine learning model was designed by integrating the hybrid marine predator algorithm and atom search optimization with an artificial neural network. The network is trained by the Gene Expression Omnibus dataset as input, and the patients' data are used as input to predict symptoms. The experimental results demonstrate that the proposed model's accuracy is higher than state-of-the-art prediction models at approximately 99.70%.

• 한국정밀공학회:학술대회논문집
• /
• 한국정밀공학회 2006년도 추계학술대회 논문집
• /
• pp.675-676
• /
• 2006

• 한국정밀공학회지
• /
• 제29권2호
• /
• pp.139-146
• /
• 2012

A multi-tasking machine tool for large scale marine engine crankshafts has been developed together with design technologies for its special devices. Since work pieces, that is, crankshafts to be machined are big and heavy; weight of over 100 tons, length of 10 m long, and diameter of over 3.5 m, several special purpose core devices are necessarily developed such as PTD (Pin Turning Device) for machining eccentric pin parts, face place and steady rest for chucking and resting heavy work pieces. PTD is a unique special purpose device of open-and-close ring typed structure equipped with revolving ring spindle for machining eccentric pins apart from journal. In order to achieve high rigidity of the machine tool, structural design optimization using TMSA (Taguch Method based Sequential Algorithm) has been completed with FEM structural analysis, and a hydrostatic bearing system for the PTD has been developed with theoretical hydrostatic analysis.

• Journal of Magnetics
• /
• 제21권4호
• /
• pp.537-543
• /
• 2016

Currently, an interest in electric vehicles (EVs) exhibited by automakers, government agencies and customers make it as more attractive research. This is due to carbon dioxide emitted by conventional combustion engine that worsens the greenhouse effect nowadays. Since electric motors are the core of EVs, it is a pressing need for researchers to develop advanced electric motors. As one of the candidates, switched flux machine (SFM) is initiated in order to cope with the requirement. This paper proposes a new alternate circumferential and radial flux (AlCiRaF) of permanent magnet switched flux machines (PMSFM) for light weight electric vehicles. Firstly, AlCiRaF PMSFM is compared with the conventional PMSFM based on some design restrictions and specifications. Then the design refinements techniques are conducted by using deterministic optimization method in order to improve preliminary performance of machine. Finally the optimized machine design has achieved maximum torque and power of 47.43 Nm and 12.85 kW, respectively, slightly better than that of conventional PMSFM.

• ETRI Journal
• /
• 제44권4호
• /
• pp.613-623
• /
• 2022

Autism spectrum disorder (ASD) is a developmental disorder associated with cognitive and neurobehavioral disorders. It affects the person's behavior and performance. Autism affects verbal and non-verbal communication in social interactions. Early screening and diagnosis of ASD are essential and helpful for early educational planning and treatment, the provision of family support, and for providing appropriate medical support for the child on time. Thus, developing automated methods for diagnosing ASD is becoming an essential need. Herein, we investigate using various machine learning methods to build predictive models for diagnosing ASD in children using facial images. To achieve this, we used an autistic children dataset containing 2936 facial images of children with autism and typical children. In application, we used classical machine learning methods, such as support vector machine and random forest. In addition to using deep-learning methods, we used a state-of-the-art method, that is, automated machine learning (AutoML). We compared the results obtained from the existing techniques. Consequently, we obtained that AutoML achieved the highest performance of approximately 96% accuracy via the Hyperpot and tree-based pipeline optimization tool optimization. Furthermore, AutoML methods enabled us to easily find the best parameter settings without any human efforts for feature engineering.

• 산업경영시스템학회지
• /
• 제16권27호
• /
• pp.73-82
• /
• 1993

In this paper, We determine the expected travel time for several forks Storage/Retrieval machine which is allowed multiple stops in aisle. When throughput is increased, We propose adding to fork number of each S/R machine rather than adding to number of S/R machine, We also describe such a model which determines the optimal number of each several forks S/R machine subject to constraints on the hourly throughput and warehouse dimensions. Numerical example is presented to compare warehouse shapes against each single fork, twin forks, triple forks S/R machine for various throughput values.