Learning Instance-Based Decision Rules for Real Time Scheduling

Rhee, Jongtae 東國大學校 1992 東國論叢 Vol.31 No.-

실시간 계획수립을 위한 규칙들을 경험적 사례들로부터 직접 획득하는 방법을 연구하였다. 이러한 사례들은 해당분야의 전문가의 의사결정행위로부터 얻어지는데 제시된 방법에 의해 새로운 패턴이 주어질 때마다 기존의 규칙들이 개선되고 또한 필요한 규칙들이 새로이 만들어지게 된다. 이 규칙들을 이용하여 의사결정을 자동화할 경우 의사결정은 사례를 제공한 전문가의 행위와 유사하게 된다. 특히 본 논문에서는 주어진 상황에서 출력변수와 가능한 값들에 대한 확률을 줄 수 있는 규칙들을 학습하고 이를 의사결정에 사용할 수 있는 방법을 개발하였다. 제시된 방법을 흐름생산라인에서 로보트의 운영문제에 적용하였다. A recursive process of learning decision rules from instance patterns to be used for real time scheduling has been developed. The Patterns are obtained from schedules made by an expert decision maker. By the process, rules are updated and if necessary, new rules are created directly from each instance pattern. When scheduling with the learned rules, given values to input elements, the set of rules will produce a decision such that it is the most consistent with the patterns from which rules have been learned. The process has been tested with a problem of scheduling material handling robot in a flow shop line.

Flow shop의 효율제고를 위한 동적 작업배정방안

이종태(Jongtae Rhee) 대한산업공학회 1994 대한산업공학회지 Vol.20 No.4

유연생산시스템의 실시간 운용을 위한 작업배정규칙의 모의실험에 의한 습득방법 : Flow shop의 경우 Flow Shop Case

이종태 동국대학교 산업기술연구원 1993 산업기술논문집 Vol.5 No.-

A simulation based methodology to obtain dispatching rules to be used for real time operation of highly automated manufacturing system, such as FMS, is suggested. That is, the dispatching rules are obtained from simulation with computer model of real manufacturing system. This approach will be useful in the case where the manufacturing system has very dynamic and complicated characteristics so that mathematical approach is almost impossible and where there is no human operation expert. The suggested method has been applied to a simple case of flow shop FMS and useful dispatching rules, which can increase the efficiency of the manufacturing system, has been obtained from simulation.

Job shop의 효율제고를 위한 동적 작업배정방안

이종태 東國大學校 1994 東國論叢 Vol.33 No.-

The productivity of manufacturing system is depending on the mean flow time and throughput rate of products. The performance in these aspects is mainly determined by real time dispatching of jobs in the case jobs are released dynamically. In this research, a dynamic dispatching method is suggested for the job shop case where a new job with random process time and process sequence is arrived with exponential interarrivel time and released into the system if the number of WIP in the system is within a predetermind limit. The proposed method has been compared with some classical dispatching methods by simulation.

자기조직화 신경망을 이용한 복수차량의 실시간 경로계획

이종태,장재진 한국경영과학회 2000 韓國經營科學會誌 Vol.25 No.4

This work proposes a neural network approach to solve vehicle routing problems which have diverse appllcation areas such as vehicle routing and robot programming. In solving these problems, classical mathematical approaches have many difficulties. In particular, it is almost impossible to implement a real-time vehicle routing with multiple vehicles. Recently, many researchers proposed methods to overcome the limitation by adopting heuristic algorithms, genetic algorithms, neural network techniques and others. The most basic model for path planning is the Traveling Salesman Problem(TSP) for a minimum distance path. We extend this for a problem with dynamic upcoming of new positions with multiple vehicles. In this paper, we propose an algorithm based on SOM(Self-Organization Map) to obtain a sub-optimal solution for a real-time vehicle routing problem. We develop a model of a generalized multiple TSP and suggest an efficient solving procedure.

Dog image Matching Using Breed Probability

Bohan Yoon,Jongtae Rhee 한국콘텐츠학회 2019 한국콘텐츠학회 ICCC 논문집 Vol.2019 No.12

다층신경망의 학습능력 향상을 위한 학습과정 및 구조설계

이대식,이종태 한국경영과학회 2001 經營 科學 Vol.18 No.2

The well-known back-propagation algorithm for multi-layered neural network has successfully been applied to pattern classification problems with remarkable flexibility. Recently, the multi-layered neural network is used as a powerful data mining tool. Nevertheless, in many cases with complex boundary of classification, the successful learning is not guaranteed and the problems of long learning time and local minimum attraction restrict the field application. In this paper, an Improved learning procedure of multi-layered neural network is proposed. The procedure is based on the generalized delta rule but it is particular in the point that the architecture of network is not fixed but enlarged during learning. That is, the number of hidden nodes or hidden layers are increased to help finding the classification boundary and such procedure is controlled by entropy evaluation. The learning speed and the pattern classification performance are analyzed and compared with the back-propagation algorithm.

RFID 기술을 이용한 u-CRM에 관한 탐색적 연구

양승정,이종태 동국대학교 산업기술연구원 2007 산업기술논문집 Vol.16 No.2

The RFID the technique is non-contact radio recognition technique and overcomes the bar codes of the limit of slow recognition speed, recognition ratio and storage capability. It will be like that and it is paid attention with the fact that it will bring a renovation change in enterprise goods activity or information transmission system. It is a focus technique that simultaneously takes charge of a ubiquitous network sensor function. The u-CRM uses a RFID technique. In ubiquitous environment, the u-CRM is to collect, to store a customer data and to manage the customer. And at real-time it will collect and will be able to analyze information of the customer. Therefore it will be able to provide an immediate service. This study, it compares the u-CRM that uses the RFID and existing CRM and it proposes the efficient u-CRM model.

Impact of IT Education on Organizational Performance in the Agricultural Sector

Jihye You,Junghoon Moon,Rhee Cheul,Jongtae Lee 대한경영정보학회 2016 경영과 정보연구 Vol.35 No.3

본 연구는 정보화 교육이 농업법인회사에 미치는 영향을 밝히는데 목적을 두었다. 이를 위해, 본 연구는 농림수산식품교육문화정보원에서 실시한 정보화수준 및 활용도 조사 보고서를 이용하여 구조방정식 모델 분석을 실시하였다. 연구결과, 정보화 교육은 소셜네트워크서비스(Social Network Services, SNS)를 활용하는 업무의 비율과 정보통신융합기술(Information and Communications Technologies, ICT)을 통해 축적된 데이터를 사용하는 업무의 비율에 긍정적인 영향을 주었다. 특히, ICT 정보시스템 활용은 조직 내 업무 효율성을 높이는 것으로 나타났다. 그리고 결국, 업무 효율성은 농업 경영 조직의 업무 효과성을 증대시키는 것을 확인하였다. 따라서 농업분야에 대한 정보화 교육은 ICT를 활용한 정보시스템 교육을 중점으로 하여 이뤄져야 할 것이다. This study aimed to clarify the effect of information technology (IT) education on the efficiency and effectiveness of working processes among agriculture corporations. Survey data on information levels from 222 agriculture corporations were collected from the Korean Agency of Education, Promotion, and Information Service in Food, Agriculture, Forestry, and Fisheries (EPIS) for a governmental white paper. Structural equation modeling was used for analysis. This study found that IT education increases the ratio of the use of information systems in working processes, especially given the use of data accumulated through information and communications technologies (ICT). The findings of this study suggest that the use of ICT data as an aspect of IT education is beneficial for the agricultural sector.

IP Multicast 트리 구현을 위한 유전자 알고리즘

민철기,이종태 동국대학교 산업기술연구원 2000 산업기술논문집 Vol.11 No.2

정보화 시대에 이르러 통신에 대한 수요가 기하 급수적으로 늘고 있고, 급속한 망의 고도화 및 확장이 일어나고 있다. 새로운 네트워크를 구축하거나, 기존의 네트워크를 확장 및 재구성 할 때 최적 네트워크 설계를 위한 방법으로 전통적인 OR 기법과 휴리스틱 방법이 쓰여 있다. 망의 구성에는 노드간의 연결비용이 발생한다. 또한 망의 각 노드는 처리용량이 제한되어 있으므로 수요량을 만족하는 망의 설계를 할 수 있어야 한다. 본 연구의 목적은 이와 같이 노드의 용량이 제한된 경우에 전체수요를 만족할 수 있는 최소비용의 망 구조를 구축하는 것이다. 네트워크 설계 문제는 NP-Complete인 경우가 대부분이며 최적해를 구할 수 없는 경우가 일반적이다. 본 연구에서는 이 문제를 효율적으로 해결하기 위하여 메타 휴리스틱 방법중 유전자 알고리듬을 이용하도록 한다. 본 연구에서 제안하는 유전자 알고리즘을 시뮬레이션에 의해 종래의 방법론에 비해 짧은 세대수 이내에 만족해를 얻을 수 있는 것으로 조사되었다. Multicast is a network-layer function that constructs paths along which data packets from a source are distributed to many, but not all, of the destinations in a communication network. For multicasting, the paths to receivers share links and the network implicitly defines a tree structure to distribute multicast packets. The problem of computing delay-constrained minimum-cost multicast trees in a network is of great interest in the last years. The best known solution to this problem is the Bounded Shortest Multicast Algorithm(BSMA), which suffers high complexity and requires very large computation time. To resolve these problems, neural networks or genetic algorithm are applied in recent researches. In this paper, a genetic algorithm to construct Multicast Tree Structure is proposed. Simulation results shows that the proposed algorithm yields efficient tree structures in the aspect of network operating cost and the quality of service than BSMA for relatively large networks.