대기행열모형에서 통제변수를 이용한 시뮬레이션 효율성 평가

권치명 東亞大學校 經營問題硏究所 1996 經營論叢 Vol.17 No.-

고객-상품 구매행렬 기반 전자상거래 추천자 시스템의 시뮬레이션 분석

권치명,김성연 한국자료분석학회 2007 Journal of the Korean Data Analysis Society Vol.9 No.4

Recommender systems for E-commerce site receive information from customers about which products they are interested in, and recommend products that are likely to fit their needs. This paper investigates the efficiencies of the collaborative filtering method and a SVD(singular value decomposition)-based recommender system for the purpose of producing useful recommendations to customers when large-scale customer-product purchase data are available.Reduced product-dimensionality obtained from SVD requires less computational effort for obtaining the neighborhood for target customer, thus it may improve the efficiency of recommendation performance. Simulation experiments on synthetic customer-product purchase data show SVD-based recommender system yields a better performance than the CF method with respect to the recall, precision and F1 measure. In applying SVD recommender system, the recommendation quality increases as the size of the neighborhood increases up to 30, but above that point, the improvement gains diminish a little. Our simulation results show an appropriate number of products for recommendation would be 10 in term of the errors of false positives. Around 10, the recall is not small, and both precision and F1 measure show good performances. We consider these informations may be useful in applying recommender system. 본 연구는 고객의 상품에 대한 평가 자료 대신 상품 구매 리스트가 활용 가능한 경우에 SVD 기법을 상품 추천 시스템에 활용하였다. SVD에 의하여 얻어진 낮은 차원의 변환 상품 구매행렬은 인접고객군을 발견하는 계산 노력을 감소시킴으로써 상품 추천 효율성을 개선할 것으로 기대한다. 가상적인 상품 구매 리스트를 대상으로 실행한 시뮬레이션 실험 결과에서 고객-상품 구매행렬의 SVD에 추천자 시스템은 CF 기법보다 recall, precision 및 F1 측도에서 우수하게 나타났다. SVD에 의한 추천자 시스템에서 목표고객에 대한 인접고객군의 크기가 30 정도 수준 이상이면 추천자 시스템의 평가 측도는 별 변화가 없는 것으로 나타나고 있다. 추천 상품 수가 증가함에 따라 추천자 시스템의 recall은 증가하고 정확도는 감소하는데 본 연구결과도 이와 유사하게 나타나고 있었으며 대략 추천 상품 수가 10개 정도일 때 recall이 적지 않으면서 추천자 시스템의 정확도가 높게 나타나고 있다. 추천자 시스템을 구현하는 용도에 따라 이러한 정보는 유용하게 사용될 수 있다고 판단된다.

통제변수를 이용한 PERT 네트워크에서 프로젝트 완료확률의 추정

권치명 한국시뮬레이션학회 2000 한국시뮬레이션학회 논문지 Vol.9 No.4

Often system analysts are interested in the estimation of percentile for system performance. For instance, in PERT network system, the percentile that the project. Typically the control variate method is used to reduce the variability of mean response using the correlation between the response and the control variates with a little additional cost during the course of simulation. In the same spirit, we apply this method to estimate the percentile of project completion time in PERT system, and evaluate the efficiency of the controlled estimator for its percentile.1 Simulation results indicate that the controlled estimators are more effective in reducing the variances of estimators than the simple estimators, however those tend to a little underestimate the percentiles for some critical values. We need more simulation experiments to examine such a kind of bias problem. We expect this research presents a step forward in the area of variance reduction techniques of stochastic simulation.

Parametric LCP를 이용한 LP해법

權致明 東亞大學校 經營問題硏究所 1985 經營論叢 Vol.8 No.-

In this paper, approach to solving the LP similiar to Dantzig's self-dual algorithm is described. By the transformation of the LP to the Parametric LCP using parameter t, we search the primal and dual feasible path. Some properties of algorithm are: 1) It solves the primal-dual problem at the same time. 2) It searchs the primal-dual feasible path. 3) It is the convergent algorithm of parameter t. 4) Choosing the initial feasible solution can effect the behavior of algorithm. 5) If the feasible start point is chosen, then the path will be primal-dual beasible throughout the algorithm. 6) At each step of algorithm, it provides a good bound of optimal objective function. In solving the large scale problem by the simplex algorithm, in many cases, 95% of optimality may be achieved. In these cases, disappropriate amount of time may be spent in obtaining very eittle improvement. There is no way of halting to do this in the simplex algorithm. Algorithm described will be appropriate for such problems.

2人 섰다 게임

權致明,朴淳達 한국경영과학회 1982 한국경영과학회지 Vol.7 No.2

‘섰다’게임은 흔히 애용되는 게임이다 이 論文은 이 섰다 게임의 模型化를 시도한 것으로써 특히 2人섰다게임을 2人零合게임(two-person zero-rum game)으로 模型化하여 最適解를 구해 보았다 이 2人섰다 게임은 先과 또 한 사람사이의 섰다게임으로 판돈과 설 때 내는 돈의 액수에 따라 最適解가 달라지는 데 예로써 판돈보다 설 때 내는 돈의 3배일때는 先은 7곳이상일 때 서는 것이 最適이고 상대방은 9끗 이상일 때 서는 것이 最適이다 이때 게임의 값은 0.35이다

다분포 대형 시뮬레이션 모형에 대한 결합상관방법

권치명 한국시뮬레이션학회 1992 한국시뮬레이션학회 논문지 Vol.1 No.1

This research develops two variance reduction methods for estimating the parameters of the experimental simulation model having multiple design points based on an approach focusing on reduction of the variances of the mean responses across multiple design points. The first method extends a combined approach of antithetic variates and control variates for a single design point to the multipopulation context with independent streams across the design points. The second method extends the same strategy in conjunction with the Schruben-Margolin method for improving the first method. We illustrate an example for implementing the second method. We expect these two approaches may improve the estimation of the parameters of interest compared with the control variates method.

중심합성계획 시뮬레이션 실험에서 공통난수의 활용

권치명,Kwon, Chi-Myung 한국시뮬레이션학회 2014 한국시뮬레이션학회 논문지 Vol.23 No.3

The central composite design (CCD) is often used to estimate the second-order linear model. This paper uses a correlation induction strategy of common random numbers (CRN) in simulation experiment and utilizes the induced correlations to obtain better estimates for the second-order linear model. This strategy assigns the CRN to all design points in the CCD. An appropriate selection of the axial points in CCD makes the weighted least squares (WLS) estimator be equivalent to ordinary least squares (OLS) estimator in estimating the linear model parameters of CCD. We analytically investigate the efficiency of this strategy in estimation of model parameters. Under certain conditions, this correlation induction strategy yields better results than independent random number strategy in estimating model parameters except intercept. The simulation experiment on a selected model supports such results. We expect a suggested random number assignment is useful in application of CCD in simulation experiments.

수요가 재생 도착과정을 따르는 (s, S) 재고 시스템에서 시뮬레이션 민감도 분석을 이용한 최적 전략

권치명 한국시뮬레이션학회 2003 한국시뮬레이션학회 논문지 Vol.12 No.3

This paper studies an optimal policy for a certain class of (s, S) inventory control systems, where the demands are characterized by the renewal arrival process. To minimize the average cost over a simulation period, we apply a stochastic optimization algorithm which uses the gradients of parameters, s and S. We obtain the gradients of objective function with respect to ordering amount S and reorder point s via a combined perturbation method. This method uses the infinitesimal perturbation analysis and the smoothed perturbation analysis alternatively according to occurrences of ordering event changes. The optimal estimates of s and S from our simulation results are quite accurate. We consider that this may be due to the estimated gradients of little noise from the regenerative system simulation, and their effect on search procedure when we apply the stochastic optimization algorithm. The directions for future study stemming from this research pertain to extension to the more general inventory system with regard to demand distribution, backlogging policy, lead time, and inter-arrival times of demands. Another direction involves the efficiency of stochastic optimization algorithm related to searching procedure for an improving point of (s, S).

0 - 1 다단계배낭기법

권치명,정성진,Gwon Chi-Myeong,Jeong Seong-Jin 한국국방경영분석학회 1984 한국국방경영분석학회지 Vol.10 No.1

The 0-1 multi-period Knapsack problem (MPKP) has a horizon of m periods, each having a number of types of projects with values and weights. Subject to the requirement, the cummulative capacity of the problem in each period i cannot be exceeded by the total weight of the projects selected in period 1, 2, ..., i. It is a problem of selecting the projects in such a way that the total value in the knapsack through the horizon of m periods is maximized. A search algorithm is developed and tested in this paper. Search rules that avoid the search of redundant partial solutions are used in the algorithm. Using the property of MPKP, a surrogate constraint concerned with the most available requirement is used in the bounding technique.

통제변수 기반 Gradient를 이용한 확률적 최적화 기법

권치명,김성연,Kwon, Chi-Myung,Kim, Seong-Yeon 한국시뮬레이션학회 2009 한국시뮬레이션학회 논문지 Vol.18 No.2

In this paper, we investigate an optimal allocation of constant service resources in stochastic system to optimize the expected performance of interest. For this purpose, we use the control variates to estimate the gradients of expected performance with respect to given resource parameters, and apply these estimated gradients in stochastic optimization algorithm to find the optimal allocation of resources. The proposed gradient estimation method is advantageous in that it uses simulation results of a single design point without increasing the number of design points in simulation experiments and does not need to describe the logical relationship among realized performance of interest and perturbations in input parameters. We consider the applications of this research to various models and extension of input parameter space as the future research.