• Journal of Korean Society of Transportation
• /
• v.25 no.6
• /
• pp.209-230
• /
• 2007

This paper analyzed the social cost function of a congestion-prone service system, which is developed from the social cost minimization problem. The analysis focused on the following two issues that have not been explicitly explored in the previous studies: the effect of the heterogeneity of value-of-travel-times among customers on the structure of cost functions; and the structure of the supplier cost function constituting the social cost function. The analysis gave a number of findings that could be summarized as follows. First, the social marginal cost for one unit increase in system output having a certain value-of-travel-time is the sum of the service time cost for that value-of-travel-time and the marginal congestion cost for the average value-of-service-time of all the system outputs. Second, the marginal congestion cost equals the marginal supplier cost of system output under the condition that supplier compensates the customers for the changed service time costs which is incurred by the marginal capacity increase necessary for economically facilitating an additional system output. Third, the compensated marginal cost is the multiple of the marginal capacity cost and the inverse of system utilization ratio, if the service time function is homogeneous of degree zero in its inputs.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.34 no.10C
• /
• pp.1001-1009
• /
• 2009

In this paper, an efficient blind frequency offset estimation method for radio resource saving in orthogonal frequency division multiplexing (OFDM) systems is proposed. In the proposed method, we obtain two time different received OFDM signal blocks by using the cyclic prefix and define the cost function by using the two OFDM signal blocks. We show that the cost function can be approximately expressed as a closed form cosine function. The approximated cosine function can be obtained from three independent cost function values calculated at three different frequency offsets. In the proposed method, the frequency offset can be estimated by calculating a frequency offset minimizing the approximated cosine function without searching all the frequency offset range. Unlike the conventional methods such as MUSIC method, the accuracy of the proposed method is independent of the searching resolution since the closed form solution exists. The computer simulation shows that the performance of the proposed method is superior to those of the MUSIC and the oversampling method.

• Journal of Digital Convergence
• /
• v.18 no.2
• /
• pp.261-268
• /
• 2020

Machine learning is achieved by making a cost function from a given digital signal data and optimizing the cost function. The cost function here has local minimums in the cost function depending on the amount of digital signal data and the structure of the neural network. These local minimums make a problem that prevents learning. Among the many ways of solving these methods, our proposed method is to change the learning step-size. Unlike existed methods using the learning rate (step-size) as a fixed constant, the use of multivariate function as the cost function prevent unnecessary machine learning and find the best way to the minimum value. Numerical experiments show that the results of the proposed method improve about 3%(88.8%→91.5%) performance using the proposed method rather than the existed methods.

• The Journal of the Acoustical Society of Korea
• /
• v.20 no.3
• /
• pp.8-14
• /
• 2001

In this paper, we present a new cost function for subband block adaptive algorithm and block conjugate gradient algorithm for noise cancellation of acoustic signal. For the cost function, we process the subband signals with data blocks for each subbands and recombine it a whole data block. After these process, the cost function has a quadratic form in adaptive filter coefficients, it guarantees the convergence of the suggested block conjugate gradient algorithm. And the block conjugate gradient algorithm which minimizes the suggested cost function has better performance than the case of full-band block conjugate gradient algorithm, the computer simulation results of noise cancellation show the efficiency of the suggested algorithm.

• Proceedings of the KOR-KST Conference
• /
• 2007.05a
• /
• pp.89-97
• /
• 2007

본 연구에서는 통행배정 모형이 갖는 여러 가지 가정 중 대칭적 통행비용 함수를 갖는 가정을 극복할 수 있는 방법에 대해 살펴보았다. 통행배정 문제에 있어서 대칭적 비용함수 가정이라는 것은 링크의 통행비용은 다른 링크의 교통량에 전혀 영향을 받지 않는 않으면서, 동시에 해당 링크를 통과하는 단하나의 수단에 의해서만 결정된다는 의미이다. 본 연구에서는 이러한 가정을 극복할 수 있는 비대칭 통행배정모형의 특성을 살펴보고, 그 해석 모형에 대해 고찰하였다. 이 때 대표적 비대칭 통행배정 문제인 다수단 통행배정 모형을 중심으로 문제를 정의하여 검토하였다. 대각화(Diagonalized) 알고리즘과 Column Generation에 기반한 heuristic 모형을 다수단 통행배정 모형에 적용하여 그 결과를 분석하였다. 그 과정을 통해 대각화 알고리즘은 초기해의 수단과 수렴기준 수단에 따라 서로다른 해를 갖는 복수의 평형해(Equilibria)특성을 가지고 있었다. 그에 비해 Column Generation에 기반한 heuristic 모형은 Euclidean Norm을 이용한 부분최적화를 통해 복수의 평형해(Equilibria)에 관한 문제점을 개선할 수 있었다.

• Journal of the Institute of Electronics Engineers of Korea SD
• /
• v.43 no.10 s.352
• /
• pp.134-139
• /
• 2006

In this paper we present the Low-Power Bus Driven Floorplan(BDF) in which the bus power consumption is minimized by using a new cost function. The previously reported BDF has used the cost function which minimizes only the chid and the bus area. However, such a cost function may not consider the bus power consumption determined by the topology of a bus in case of the segmented bus design. In this paper, we formulate a new cost function which. reflects the communication frequency and the real distance between blocks in a bus to model the bus power consumption. For the Low-Power BDF with the new cost function, the experimental results show the bus power consumption cost is reduced by 11.43% on the average.

• Environmental and Resource Economics Review
• /
• v.16 no.1
• /
• pp.171-197
• /
• 2007

In this paper we provide a methodology that permits test of feasibility of Porter hypothesis under limited data environment by utilizing two types of output distance function: Shephard distance function and directional distance function. The production technology supported by Porter hypothesis is embodied in the directional distance function. The average annual marginal abatement cost for $SO_2$ obtained by estimating the directional distance function is higher than the one obtained with the Shephard distance function by about 50 percent in the Korean electric power industry over the period 1970~1998. This result implies that feasibility of Porter hypothesis depends upon whether investment in production capital and process can bring more than a 50 percent increase in the average productivity.

• Proceedings of the Korean Reliability Society Conference
• /
• 2000.04a
• /
• pp.55-61
• /
• 2000

본 연구에서는 열화시스템의 교체정책을 결정할 때 열화에 의한 시스템의 성능의 저하로 인한 제품의 품질 특성치와 목표치(target value)와의 차이를 손실함수를 통해 비용으로 반영하여 교체정책을 제시하였다. 비용함수에는 검사비용, 교체비용, 손실비용이 포함되며, 이 비용들을 최소화하는 교체정책이 제시되었다. 또한 교체정책 결정 시 고려되는 열화에 의한 특성치의 변화, 검사비용, 열화추세 등의 변화에 따른 총비용의 변화와 상호관계를 실험을 통해서 관찰하였고, 실험결과 시스템의 특성에 따라 총비용을 최소화하는 최적검사주기가 존재함을 알 수 있다. 열화에 의한 특성치의 변화가 교체주기와 비용의 변화에 크게 영향을 주는 것으로 나타난 반면 검사비용은 정책결정에 큰 영향을 주지 않는 것으로 관측되었다. 손실함수의 도입을 통해 품질 특성치와 목표치와의 차이로 인한 비용을 줄일 수 있으며 이는 교체주기의 감소경향을 나타낸다.

• Environmental and Resource Economics Review
• /
• v.19 no.3
• /
• pp.633-658
• /
• 2010

In this paper, we estimate the Translog cost function in Korean manufacturing, using capital (K), labor (L), material (M), electricity (E), fuel (F) data over the period from 1970 to 2005. Especially, this paper investigates the impact of imposing concavity in the estimation of a Translog cost function. Although the value of log-likelihood is somewhat reduced in a concavity imposed function rather than a function which is not, a concavity imposed function satisfies regularity conditions (monotonicity, positivity, concavity) at all data points. We also calculate price elasticities using a concavity imposed Translog cost function. Electricity complements capital so electricity demand increases as capital demand increases. Meanwhile, electricity substitutes labor, fuel, and material. These results show that Korean manufacturing experienced a structural change of increase in electricity demand.

• The KIPS Transactions:PartD
• /
• v.15D no.2
• /
• pp.207-212
• /
• 2008

This paper evaluates software cost estimation models, and presents the most suitable model. First, we transformed a relevant model into variables to make in linear. Second, we evaluated model's performance considering how much suitable the cost data of the actual development software was. In the stage of model performance evaluation criteria, we used MMRE which is the relative error concept rather than the absolute error. Existing software cost estimation model follows Weibull, Gamma, and Rayleigh function. In this paper, Gompertz function model is suggested which is a kind of growth curve. Additionally, we verify the compatability of other different growth curves. As a result of evaluation of model's performance, Gompertz function was considered to be the most suitable for the cost estimation model.