• 대한전자공학회:학술대회논문집
• /
• 대한전자공학회 2002년도 하계종합학술대회 논문집(3)
• /
• pp.135-138
• /
• 2002

This paper addresses incremental learning method for regression. SVM(support vector machine) is a recently proposed learning method. In general training a support vector machine requires solving a QP (quadratic programing) problem. For very large dataset or incremental dataset, solving QP problems may be inconvenient. So this paper presents an incremental support vector learning method for function approximation problems.

• 대한수학회보
• /
• 제53권2호
• /
• pp.589-600
• /
• 2016

In this paper, we propose a new incremental gradient method for solving a regularized minimization problem whose objective is the sum of m smooth functions and a (possibly nonsmooth) convex function. This method uses an adaptive stepsize. Recently proposed incremental gradient methods for a regularized minimization problem need O(mn) storage, where n is the number of variables. This is the drawback of them. But, the proposed new incremental gradient method requires only O(n) storage.

• Industrial Engineering and Management Systems
• /
• 제8권3호
• /
• pp.194-200
• /
• 2009

This study provides a heuristic algorithm to solve the locating problem of repair service centers (RSCs). To enhance the customer service level with more satisfaction and quicker responsiveness, the locating problem of RSCs has become one of the important issues in reverse supply chain management. This problem is formulated as a zero-one mixed integer programming in which an exiting distributor will be considered to be an un-capacitated repair service center for the objective of cost-minimizing. Since logistical costs are highly interrelated with the multinational location of distributors and RSCs, the fixed cost for setting a repair service center, variable cost, transportation cost, and exchange rates are considered in this study. Recognizing the selection of un-capacitated RSCs' locations is a combinatorial optimization problem and is a zero-one mixed integer programming with NP-hard complexity, we provide a heuristic algorithm named as incremental cost matrix procedure (ICMP) to simplify the solving procedure. By using the concise and structural cost matrix, ICMP can efficiently screen the potential location with cost advantage and effectively decide which distributor should be a RSC. Results obtained from the numerical experiments conducted in small scale problem have shown the fact that ICMP is an effective and efficient heuristic algorithm for solving the RSCs locating problem. In the future, using the extended ICMP to solve problems with larger industrial scale or problems with congestion effects caused by the variation of customer demand and the restriction of the RSC capacity is worth a further investigation.

• 한국정보처리학회논문지
• /
• 제3권3호
• /
• pp.429-438
• /
• 1996

본 논문에서는 주어진 문제를 해결하기 위해 사용될 최적의 다충 퍼센트론을 구성 하기 위한 하나의 점진적 학습 방법을 제시한다. 고정된 크기의 트레이닝 패턴 집합을 반복적으로 사용하는 기존의 알고리즘들과는 달리, 제시되는 방법에서는 학습 패턴의 수를 점차 증가시키면서 전체 데이터를 학습하기 위해 필요하고도 충분한 은닉뉴런의 수를 찾는다. 이와 같이 신경망 크기의 최적화에 학습 패턴을 점차적으로 선택하여 늘려나감으로써 일반화 능력과 학습 속도가 기존의 방법에서보다 향상됨을 필기체 숫자인식 문제에 있어서 실험적으로 보여준다.

• 컴퓨터교육학회논문지
• /
• 제23권3호
• /
• pp.41-47
• /
• 2020

21세기 미래 인재에게 필요한 핵심 역량으로 컴퓨팅 사고력이 주목받고 있다. 국내외적으로 컴퓨팅 사고력 향상을 위한 소프트웨어 교육이 한창이다. 그중에서 문제해결 프로그래밍 교육은 컴퓨팅 사고력 향상에 도움이 된다. CT-TDPS 학습 모형은 복잡한 문제들을 모듈화하는 분해, 추상화 사고 과정과 이를 구현하는 반복적·점증적 프로그래밍 방식인 애자일(Agile) 개발 방식을 따른다. 본 연구에서는 스크래치를 이용한 문제해결 프로그래밍 교육에 CT-TDPS 학습 모형을 적용하여 컴퓨팅 사고력 향상을 확인하고자 하였다. 연구 결과, CT-TDPS 학습 모형을 적용한 문제해결 프로그래밍 교육에서 컴퓨팅 사고력의 하위 요인인 컴퓨팅 개념, 컴퓨팅 수행, 컴퓨팅 관점에서 모두 향상이 되었음을 확인할 수 있었다. 그리고, Dr.Scratch 자동 평가 결과에 대한 t 검정 결과 실험집단에서 유의한 차이가 있음을 확인하였다.

• 경영과정보연구
• /
• 제21권
• /
• pp.49-76
• /
• 2007

The main purpose of this study is to assess the relationship between communication satisfaction and group creativity. To achieve this purpose, this study has two objectives. First, I will identify the relationship between the satisfaction of communication and idea creation. Second, the relationship between communication satisfaction and problem solving activity will be statistically tested. The major findings of the study can be summarized as follows: First, Statistically significant positive relations were found between the communication satisfaction and idea creation. Especially, sub-construct for supervisor communication$({\beta}=0.14,\;p<.05)$, quality of media$({\beta}=0.13,\;p<.05)$, individual feedback$({\beta}=0.36,\;p<.001)$, organizational prospect$({\beta}0.31,\;p<.001)$ showed the statistically significant positive(+) relationship with construct in idea creation. Second, the measurement of communication satisfaction had showed statistically significant relationship with the problem solving activity. Especially, communication with supervisor$({\beta}=0.21,\;p<.05)$, quality of media$({\beta}=0.21,\;p<.001)$, organizational prospects$({\beta}=0.22,\;p<.001)$, and communication with subordinate$({\beta}=0.37,\;p<.001)$ showed the significant positive(+) relationship with problem solving activity construct. These consequences suggest that communication satisfaction of employee is effective in the improving the group creativity in organization. In the conclusion, this study present the satisfaction factor of employee communication as a incremental factor of group creativity.

• 대한산업공학회지
• /
• 제12권2호
• /
• pp.77-87
• /
• 1986

This paper presents a heuristic method for solving the discrete-time ordering problem with quantity discounts and deterministic, time-varying demand. This algorithm utilizes a variation of the incremental cost approach(ICA) to determine a near optimal solution. The ICA is the method which reduces the total cost with reduction of the number of orders by one. In order to reduce the number of orders, if the incremental cost for one of the periods is negative, the demand of the period should be purchased in its immediate preceding period. In order to test the performance of this algorithm, an experiment is conducted that involves a large number of test problems covering a wide variety of situations. The result of the experiment shows that the proposed algorithm has 80.5% better solutions than the adjusted part period algorithm(APPA), which is known to be the best heuristic method.

• 산업경영시스템학회지
• /
• 제35권4호
• /
• pp.142-148
• /
• 2012

Release planning for incremental software development is to select and assign features in sequence of releases along a specified planning horizon. It includes the technical precedence inherent in the features, the conflicting priorities as determined by the representative stakeholders, and the balance between required and available resources. The complexity of this consideration is getting more complicated when planning releases in software product lines. The problem is formulated as a precedence-constrained multiple 0-1 knapsack problem. In this research a genetic algorithm is developed for solving the release planning problems in software product lines as well as tests for the proposed solution methodology are conducted using data generated randomly.

• 대한기계학회논문집
• /
• 제18권8호
• /
• pp.2026-2038
• /
• 1994

A numerical scheme is developed for the analysis of incipient sliding contact with orthotropic friction condition subjected to tangential load and twisting moment. The inherent nonlinearity in the orthotropic friction law has been treated by a polyhedral friction law. Then, a three-dimensional linear complementarity problem(LCP) formulation in an incremental form is obtained, and the existence of a solution is investigated. A Lemke's complementary pivoting algorithm is used for solving the LCP. The scheme is illustrated by spherical contact problems, and the effects of eccentricity of elliptical friction domain on the traction and stick region are discussed.

• Journal of Electrical Engineering and Technology
• /
• 제3권1호
• /
• pp.52-59
• /
• 2008

This paper describes the secant method for solving the economic dispatch (ED) problem with generator constraints and transmission losses. The ED problem is an important optimization problem in the economic operation of a power system. The proposed algorithm involves selection of minimum and maximum incremental costs (lambda values) and then the evaluation of optimal lambda at required power demand is done by secant method. The proposed algorithm has been tested on a power system having 6, 15, and 40 generating units. Studies have been made on the proposed method to solve the ED problem by taking 120 and 200 units with generator constraints. Simulation results of the proposed approach were compared in terms of solution quality, convergence characteristics, and computation efficiency with conventional methods such as lambda iterative method, heuristic methods such as genetic algorithm, and meta-heuristic methods like particle swarm optimization. It is observed from different case studies that the proposed method provides qualitative solutions with less computational time compared to various methods available in the literature.