Most of the open-source decision tree algorithms are based on three splitting criteria (Entropy, Gini Index, and Gain Ratio). Therefore, the advantages and disadvantages of these three popular algorithms need to be studied more thoroughly. Comparisons of the three algorithms were mainly performed with respect to the predictive performance. In this work, we conducted a comparative experiment on the splitting criteria of three decision trees, focusing on their interpretability. Depth, homogeneity, coverage, lift, and stability were used as indicators for measuring interpretability. To measure the stability of decision trees, we present a measure of the stability of the root node and the stability of the dominating rules based on a measure of the similarity of trees. Based on 10 data collected from UCI and Kaggle, we compare the interpretability of DT (Decision Tree) algorithms based on three splitting criteria. The results show that the GR (Gain Ratio) branch-based DT algorithm performs well in terms of lift and homogeneity, while the GINI (Gini Index) and ENT (Entropy) branch-based DT algorithms performs well in terms of coverage. With respect to stability, considering both the similarity of the dominating rule or the similarity of the root node, the DT algorithm according to the ENT splitting criterion shows the best results.

This study pursues to solve a batch of nonlinear parameter estimation (NPE) problems where a model interpreting the independent and the dependent variables is given and fixed but corresponding data sets vary. Specifically, we assume that the model does not have an explicit form and the discrepancy between a value from a data set and a corresponding value from the model is unknown. Due to the complexity of the problem, one may prefer to use heuristic algorithms rather than gradient-based algorithms, but the performance of the heuristic algorithms depends on their initial settings. In this study, we suggest two schemes to improve the performance of heuristic algorithms to solve the target problem. Most of all, we apply a Bayesian optimization to find the best parameters of the heuristic algorithm for solving the first NPE problem of the batch and apply the parameters of the heuristic algorithm for solving other NPE problems. Besides, we save a list of simulation outputs obtained from the Bayesian optimization and then use the list to construct the initial population set of the heuristic algorithm. The suggested schemes were tested in two simulation studies and were applied to a real example of measuring the critical dimensions of a 2-dimensional high-aspect-ratio structure of a wafer in semiconductor manufacturing.

When considering military operations that require rapid response time, forward supply operation of various type of ammunition is essential. Also, t is necessary to supply ammunition in a timely manner before an ammunition shortage situation occurs. In this study, we propose a mathematical model for allocation of ammunition to ammunition storehouse at the Ammunition Supply Post (ASP). The model has several objectives. First, it ensures that the frequent used ammunition is stored in a distributed manner at a high workability ammunition storehouses. Second, infrequent used ammunition is required to be stored intensively at a single storehouse as much as possible. Third, capacity of the storehouse and compatible storage restriction required to be obeyed. Lastly, criticality of ammunition should be considered to ensure safety distance. We propose an algorithm to find the pareto-based optimal solution using the mathematical model in a reasonable computation time. The computational results show that the suggested model and algorithm can solve the real operational scale of the allocation problem.

This study studied a system that can redesign the production site layout and respond with dynamic simulation through fabric production process innovation for smart factory promotion and digital-oriented decision making of the production process. We propose to reflect the required throughput and throughput per unit facility of fabric production process as probability distribution, and to construct data-driven metabolism such as data collection, data conversion processing, data rake generation, production site monitoring and simulation utilization. In this study, we demonstrate digital-centric field decision smartization through architectural design for the smartization of fabric production plants and dynamic simulations that reflect it.

A60 급 갑판 관통 관은 선박과 해양플랜트에서 화재사고가 발생할 경우 화염의 확산을 방지하고 인명을 보호하기 위해 수평구조에 설치되는 방화장치이다. 본 연구에서는 다양한 대리모델과 다중 섬유전자 알고리즘을 이용하여 A60 급 갑판 관통 관의 방화설계에 대한 이산변수 근사최적화를 수행하였다. A60 급 갑판 관통 관의 방화설계는 과도 열전달해석을 통해 평가하였다. 근사최적화에서 관통 관의 길이, 지름, 재질, 그리고 단열재의 밀도는 이산설계변수로 적용하였고, 제한조건은 온도, 생산성 및 가격을 고려하였다. 대리모델 기반의 근사최적설계 문제는 제한조건을 만족하면서 A60 급 갑판 관통 관의 중량을 최소화할 수 있는 이산설계변수를 결정하도록 정식화 하였다. 반응표면모델, 크리깅, 그리고 방사기저함수 신경망과 같은 다양한 대리모델이 근사최적화에 사용되었다. 근사최적화의 정확도를 검토하기 위해 최적해의 결과는 실제 계산 결과와 비교하였다. 근사최적화에 사용된 대리모델 중 방사기저함수 신경망 모델이 A60 급 갑판 관통 관의 방화설계에 대해 가장 정확한 최적설계 결과를 나타내었다.

어선과 같은 소형선박의 충돌사고는 큰 인명피해를 초래한다. 본 연구 이전에 선행된 연구에서는 충돌위험을 판단하고 경보를 발생시키는 소형선박 충돌예방 알고리즘을 개발하였다. 하지만 충돌경보와 같이 안전을 위해 제공되는 서비스는 위험을 예방할 뿐 아니라 이용자의 만족도 또한 어느 정도 수반되어야 효과적으로 기능할 수 있다. 본 연구에서는 소형선박 충돌예방 알고리즘의 실용성 향상을 위해 알고리즘을 개선하고, 알고리즘을 실제로 적용하여 개선결과 및 효과를 확인하고자 하였다. 충돌경보 서비스를 사용한 소형선박 운항자들을 대상으로 설문조사를 수행한 결과, 충돌경보의 정확도 향상과 경보 횟수 및 음량에 대한 사용자들의 요구사항이 확인되었다. 이에 따라 본 연구에서는 사용자 만족도 향상을 위해 알고리즘을 개선하였으며 실제 해상환경에서 개선된 알고리즘을 적용한 실선실험 을 수행하였다. 그 결과 개선 전보다 경보 발생 빈도가 감소하였지만, 위험한 상황에서는 경보가 비교적 꾸준히 발생하였으며, 충돌경보의 정확성과 실용성이 향상된 것으로 분석되었다. 추후에는 개선된 알고리즘의 적정성에 대한 근거자료를 마련하여 알고리즘의 실용성과 신뢰성을 확보한다면 소형선박 충돌사고 예방에 효과적으로 기여할 수 있을 것이다.

In this study, a welding heat source model was presented and verified during fiber laser welding. The multi-layered heat source model is a model that can cover most of existing studies and can be defined with a simple formula. It consists of a total of 12 parameters, and an optimization algorithm was used to find them. As optimization algorithms, adaptive simulated annealing, multi island genetic algorithm, and Hooke-Jeeves technique were applied for comparative analysis. The parameters were found by comparing the temperature distribution when the STS304L was bead on plate welding and the temperature distribution derived through finite element analysis, and all three models were able to derive a model with similar trends. However, there was a deviation between parameters, which was attributed to the many variables. It is expected that a more clear welding heat source model can be derived in subsequent studies by giving a guide to the relational expression and range between variables and increasing the temperature measurement point, which is the target value.

Welding is the most widely used technology for manufacturing in the automobile, and shipbuilding industries. Fiber laser welding is rapidly introduced into the field to minimize welding distortion and fast welding speed. Although it is advantageous to use finite element analysis to predict welding distortion and find optimized welding conditions, there are various heat source model for fiber laser welding. In this study, a welding heat source was proposed using a multi-layered heat source model that encompasses most of the existing various welding heat source models: conical shape, curved model, exponential model, conical-cylindrical model, and conical-conical model. A case study was performed through finite element analysis using the radius of each layer and the ratio of heat energy of the layer as variables, and the variables were found by comparing them with the actual experimental results. For case study, by applying Adaptive simulated annealing, one of the global optimization algorithms, we were able to find the heat source model more efficiently.

PURPOSES : The driver's ability to make a commitment has resulted in excessive force and a lack of commitment. To solve this problem, we are developing an algorithm that analyzes resolution in real-time by introducing IoT and informs drivers of the completion of compaction. METHODS : Real-time compaction was analyzed by installing accelerometers on the rollers. To evaluate the algorithms, we conducted an apparent density test. RESULTS : The algorithm data and apparent density test data showed similar trends. This means that the proposed algorithms are sufficiently reliable. However, a lack of data samples and the fact that only data prior to completion of the commitment were analyzed may indicate a lack of reliability. CONCLUSIONS : In subsequent studies, the number of samples will be increased and the data after completion of the commitment analyzed to increase reliability. Introducing a tachometer will prevent the TVL from falling sharply when the direction of the rollers' progress changes. In addition, it is also planned to upgrade the algorithms by researching cases in which the algorithms did not produce satisfactory results owing to problems such as temperature and speed.

제4차 산업혁명이 시작되면서 주목받기 시작한 인공지능은 선박에도 적용되 어 자율운항선박 개발이 진행되고 있다. 자율운항선박은 선원의 승선유무와 원 격조종자의 조종여부에 따라 총 4단계로 구분되며, 완전자율운항선박은 인간의 개입 없이 설계자가 설계한 알고리즘에 따라 최적의 결정을 내리게 된다. 그러 나 인공지능의 자율성과 예측곤란성이라는 특징으로 인하여 완전자율운항선박 의 결정이 예측하기 어려우며 이러한 결정이 항상 윤리적이라고 기대하기 어렵 다. 그러나 인공지능의 예측불가능성을 인공지능의 자율성에 의존해서는 안되므로 인간의 가치가 반영된 인공지능의 알고리즘이 설계되어야 한다. 특히, 충 돌의 위험을 피할 수 없는 상황에서 인공지능이 어떠한 원칙과 기준에 따라 결 정을 내리는 것인지 이러한 과정을 결정하는 사고 알고리즘에 대해 설명가능하 고 이러한 결정이 사회 구성원들에게 합리적으로 받아들여져야 한다. 이와 같은 인공지능의 윤리문제에 대한 논의는 자율주행자동차 분야에서 가 장 활발하게 이루어지고 있다. 그러나 인간의 가치가 일관성과 보편성을 가지 기 어렵다는 한계가 있다. 특히, 윤리적 딜레마 상황에서는 ‘실천이성의 이중성’ 으로 인하여 이성적이며 윤리적인 판단과는 다른 결정을 내리게 되며, 윤리적 이라는 행동에 대한 기준 역시 지역·문화·경제 상황에 따라 다르게 인식하고 있다. 그 결과 인공지능의 윤리적 가이드라인의 필요성에 대해 인식하면서도 사고 알고리즘에 대한 구체적인 기준을 정립하는 것에 대해서는 회의적이다. 자율주행자동차의 윤리적 딜레마 상황을 예견하여 사용과정에서 발생하게 될 현실적인 문제를 완벽하게 대응할 수 없다는 것이다. 그러나 문제가 발생한 이 후 대응책을 마련한다는 것은 현실적으로 발생한 피해를 구제하기 위한 조치일 뿐 근본적인 문제를 해결하는 방법이 아니다. 특히, 전 세계 해역을 운항하는 자율운항선박은 전 세계 구성원들에게 합리적으로 받아들여지는 결정을 내려 야 한다. 따라서 국제해사기구를 중심으로 자율운항선박과 관계된 사고 알고리 즘을 포함한 윤리적 가이드라인을 마련하고 이를 개별국가에서 정책적으로 반 영함으로써 자율운항선박 개발 단계부터 사용단계에 이르는 전 과정에서 프로 그램 개발자와 이해관계자가 윤리규범을 준수하고, 축적된 정보를 기초로 보완 작업을 통해 자율운항선박의 윤리문제에 대응해 나가야 할 것이다.

Environmental fundamental facilities have different odor emission characteristics depending on the type of treatment facilities. To overcome the limitations of the olfactometry method, research needs to be conducted on how to calculate the dilution factor from the individual odor concentrations. The aim of this study was to determine the air dilution factor estimated from manually measured concentration data of individual odor substances (22 specified odor species) in three environmental treatment facilities. In order to calculate the optimum algorism for each environmental fundamental facility, three types of facilities were selected, the concentration of odor substances in the exhaust gas was measured, and the contribution of the overall dilution factor was evaluated. To estimate the dilution factor, four to six algorism were induced and evaluated by correlation analysis between substance concentration and complex odor data. Dilution factors from O municipal water treatment (MWT) and Y livestock wastewater treatment (LWT) facilities showed high level of dilution factors, because concentration levels of hydrogen sulfide and methylmercaptan, which had low odor threshold concentrations, were high. In S food waste treatment (FWT) facility, the aldehyde group strongly influenced dilution the factor (dominant substance: acetaldehyde, i-valeraldhyde and methylmercaptan). In the evaluation of four to six algorism to estimate the dilution factor, the vector algorism (described in the text) was optimum for O MWT and Y LWT, while the algorism using the sum of the top-three dominant substances showed the best outcome for S FWT. In further studies, estimation of the dilution factor from simultaneously monitored data by odor sensors will be developed and integrated with the results in this study.

Control performance of a smart tuned mass damper (TMD) mainly depends on control algorithms. A lot of control strategies have been proposed for semi-active control devices. Recently, machine learning begins to be applied to development of vibration control algorithm. In this study, a reinforcement learning among machine learning techniques was employed to develop a semi-active control algorithm for a smart TMD. The smart TMD was composed of magnetorheological damper in this study. For this purpose, an 11-story building structure with a smart TMD was selected to construct a reinforcement learning environment. A time history analysis of the example structure subject to earthquake excitation was conducted in the reinforcement learning procedure. Deep Q-network (DQN) among various reinforcement learning algorithms was used to make a learning agent. The command voltage sent to the MR damper is determined by the action produced by the DQN. Parametric studies on hyper-parameters of DQN were performed by numerical simulations. After appropriate training iteration of the DQN model with proper hyper-parameters, the DQN model for control of seismic responses of the example structure with smart TMD was developed. The developed DQN model can effectively control smart TMD to reduce seismic responses of the example structure.

본 연구는 선박용 공기압축기의 상태기반보전 시스템에 필요한 이상치 탐지 알고리즘 적용에 대한 실험적 연구로서 고장모사 실험을 통해 시계열 전류 센서 데이터를 이용한 이상탐지 적용 가능성을 확인하였다. 고장 유형 10개에 대해 실험실 규모의 고장 모사 실험을 수행하여 정상 운전데이터와 고장 데이터를 구축하였다. 실험 결과 구축된 이상탐지 모델은 시계열 데이터의 주기에 변화를 유발하는 이상은 잘 탐지하는 반면 미세한 부하 변동에 대한 탐지 성능은 떨어졌다. 또한 오토인코더를 이용한 시계열 이상탐지 모델은 입력 시 퀀스의 길이와 초모수 조정에 따라 이상 탐지 성능이 상이한 것으로 나타났다.

본 연구는 화재진압 및 피난활동을 지원하는 딥러닝 기반의 알고리즘 개발에 관한 기초 연구로 선박 화재 시 연기감지기가 작동하기 전에 검출된 연기 데이터를 분석 및 활용하여 원격지까지 연기가 확산 되기 전에 연기 확산거리를 예측하는 것이 목적이다. 다음과 같은 절차에 따라 제안 알고리즘을 검토하였다. 첫 번째 단계로, 딥러닝 기반 객체 검출 알고리즘인 YOLO(You Only Look Once)모델에 화재시뮬레이션을 통하여 얻은 연기 영상을 적용하여 학습을 진행하였다. 학습된 YOLO모델의 mAP(mean Average Precision)은 98.71%로 측정되었으며, 9 FPS(Frames Per Second)의 처리 속도로 연기를 검출하였다. 두 번째 단계로 YOLO로부터 연기 형상이 추출된 경계 상자의 좌표값을 통해 연기 확산거리를 추정하였으며 이를 시계열 예측 알고리즘인 LSTM(Long Short-Term Memory)에 적용하여 학습을 진행하였다. 그 결과, 화재시뮬레이션으로부터 얻은 Fast 화재의 연기영상에서 경계 상자의 좌표값으로부터 추정한 화재발생~30초까지의 연기 확산거리 데이터를 LSTM 학습모델에 입력하여 31초~90초까지의 연기 확산거리 데이터를 예측하였다. 그리고 추정한 연기 확산거리와 예측한 연기 확산거리의 평균제곱근 오차는 2.74로 나타났다.

PURPOSES : This study verifies the appropriateness of the observed traffic volume using car navigation traffic volume data. METHODS : In this study, we developed an annual average daily traffic (AADT) estimation model that can verify the total amount of traffic by using navigation traffic volume data. In addition, a method to verify the appropriateness of the observed traffic volume was developed using time-based navigation traffic volume data that can check the characteristics of traffic volume at each point. RESULTS : As a result of the analysis of this study, it was found that 674 of the 697 short-duration survey spots of the freeways were appropriate and that 23 spots needed to be revised. CONCLUSIONS : As a result of the analysis of this study, it was found that there was a strong positive correlation between the observed traffic volume and the car navigation traffic volume. Thus, the appropriateness of the observed traffic was determined by verifying the total amount of observed traffic and the observed traffic volume by time.

Many of companies have made significant improvements for globalization and competitive business environment The supply chain management has received many attentions in the area of that business environment. The purpose of this study is to generate realistic production and distribution planning in the supply chain network. The planning model determines the best schedule using operation sequences and routing to deliver. To solve the problem a hybrid approach involving a genetic algorithm (GA) and computer simulation is proposed. This proposed approach is for: (1) selecting the best machine for each operation, (2) deciding the sequence of operation to product and route to deliver, and (3) minimizing the completion time for each order. This study developed mathematical model for production, distribution, production-distribution and proposed GA-Simulation solution procedure. The results of computational experiments for a simple example of the supply chain network are given and discussed to validate the proposed approach. It has been shown that the hybrid approach is powerful for complex production and distribution planning in the manufacturing supply chain network. The proposed approach can be used to generate realistic production and distribution planning considering stochastic natures in the actual supply chain and support decision making for companies.

Recently, a multi facility, multi product and multi period industrial problem has been widely investigated in Supply Chain Network(SCN). One of keys issues in the current SCN research area involves minimizing both production and distribution costs. This study deals with finding an optimal solution for minimizing the total cost of production and distribution problems in supply chain network. First, we presented an integrated mathematical model that satisfies the minimum cost in the supply chain. To solve the presented mathematical model, we used a genetic algorithm with an excellent searching ability for complicated solution space. To represent the given model effectively, the matrix based real-number coding schema is used. The difference rate of the objective function value for the termination condition is applied. Computational experimental results show that the real size problems we encountered can be solved within a reasonable time.

In recent years, importance of blockchain systems has been grown after success of bitcoin. Distributed consensus algorithm is used to achieve an agreement, which means the same information is recorded in all nodes participating in blockchain network. Various algorithms were suggested to resolve blockchain trilemma, which refers conflict between decentralization, scalability, security. An algorithm based on Byzantine Agreement among Decentralized Agents (BADA) were designed for the same manner, and it used limited committee that enables an efficient consensus among considerable number of nodes. In addition, election of committee based on Proof-of-Nonce guarantees decentralization and security. In spite of such prominence, application of BADA in actual blockchain system requires further researches about performance and essential features affecting on the performance. However, performance assessment committed in real systems takes a long time and costs a great deal of budget. Based on this motivation, we designed and implemented a simulator for measuring performance of BADA. Specifically, we defined a simulation framework including three components named simulator Command Line Interface, transaction generator, BADA nodes. Furthermore, we carried out response surface analysis for revealing latent relationship between performance measure and design parameters. By using obtained response surface models, we could find an optimal configuration of design parameters for achieving a given desirable performance level.