• 대한산업공학회지
• /
• 제28권1호
• /
• pp.112-127
• /
• 2002

A useful pattern is a pattern that contributes much to learning. For a classification problem those patterns near the class boundary surfaces carry more information to the classifier. For a regression problem the ones near the estimated surface carry more information. In both cases, the usefulness is defined only for those patterns either without error or with negligible error. Using only the useful patterns gives several benefits. First, computational complexity in memory and time for learning is decreased. Second, overfitting is avoided even when the learner is over-sized. Third, learning results in more stable learners. In this paper, we propose a pattern 'utility index' that measures the utility of an individual pattern. The utility index is based on the bias and variance of a pattern trained by a network ensemble. In classification, the pattern with a low bias and a high variance gets a high score. In regression, on the other hand, the one with a low bias and a low variance gets a high score. Based on the distribution of the utility index, the original training set is divided into a high-score group and a low-score group. Only the high-score group is then used for training. The proposed method is tested on synthetic and real-world benchmark datasets. The proposed approach gives a better or at least similar performance.

• 대한전기학회논문지:전력기술부문A
• /
• 제54권6호
• /
• pp.306-314
• /
• 2005

In this paper, a new approach that is based on EEAC & only with network solutions for CS&S in the transient stability assessment is developed. The proposed CS&S algorithm in conjunction with EEAC to include the capability of performing on-line TSA without TDS is used to calculate the critical clearing time for stability index. In this algorithm, all generators are represented by classical models and all loads are represented by constant impedance load models. The accelerating & synchronizing power coefficient as an index is determined at its disturbance through solving network equation directly. As mentioned above, a new index for generator is generally used to determine the critical generators group. The generator rotor angle is fixed for non-critical generators group, but has equal angle increments for critical generators group. Finally, the critical clearing time is calculated from the power-angle relationship of equivalent OMIB system. The proposed CS&S algorithm currently being implemented is applied to the KEPCO system. The CS&S result was remarkably similar to TSAT program and SIME. Therefore, it was found to be suitable for a fast & highly efficient CS&S algorithm in TSA. The time of CS&S for the 139 contingencies using proposed CS&S algorithm takes less than 3 seconds on Pentium 4, 3GHz Desktop.

• Asian-Australasian Journal of Animal Sciences
• /
• 제9권2호
• /
• pp.149-153
• /
• 1996

In the first step, birth weights and weights at 9 and 12 months were analysed and their variability were studied. Birth weight of kids had large within breed variation ranging from 0.50 to 1.60 kg. There were more than 100 kids with birth weights ranging from 1.2 to 1.6 kg. The birth weight of kids increased with the increase of kidding weight of dams. The weights of kids at 9 and 12 months also showed large within variation ranging from 5.0 to 18.0 kg and 16.0 to 19.0 kg respectively. In the second step, these existing large within breed variations were exploited selecting best animals for further mating. The birth weight of selected group($1.13{\pm}0.2kg$) was significantly(p<0.01) higher than the random group($0.93{\pm}0.05kg$). The weights in selected group for 3 and 6 months were $4.94{\pm}0.04$ and $8.40{\pm}0.05kg$ and that of random group was $3.99 {\pm}0.08$ and $7.07{\pm}0.11kg$ respectively and differences wee significant. Significantly higher average daily gains were noticed for selected group ($42.7{\pm}0.43g/day$) as against random group($33.3{\pm}0.81g/day$). From the results it may be concluded that selection and mating of best performing animals have some effect on the improvement of growth performance of kids which may be exploited for future planning and development in the field of goat production.

• 한국물환경학회지
• /
• 제25권6호
• /
• pp.886-895
• /
• 2009

The guideline of selection of Integrated Management Practices (IMPs), such as wood, green roof, lawn, and porous pavement, for Low Impact Development (LID) design was proposed by ranking the reduction rate of surface runoff using LIDMOD1.0. Based on the guideline, LID was designed with several scenarios at two apartment complexes located at Songpa-gu, Seoul, Korea, and the effect of LID on surface runoff was evaluated during last 10 years. The effect of runoff reduction of IMP by land use change was highly dependent on the kind of hydrologic soil group. The wood planting is the best IMPs for reduction of surfac runoff for all hydrologic soil groups. Lawn planting is an excellent IMP for hydrologic soil group A, but reduction rate is low where soil doesn't effectively drains precipitation. The green roof shows constant reduction rate of surface runoff because it is not influenced by hydrologic soil group. Compared to the rate of other IMPs, the green roof is less effect the surface runoff reduction for hydrologic soil group A and is more effect for hydrologic soil group C and D followed to planing wood. The porous pavement for the impervious area is IMPs which is last selected for LID design because of the lowest reduction rate for all hydrologic soil group. As a result of LID application at study areas, we could conclude that the first step of the strategy of LID design at apartment complex is precuring pervious land as many area as possible, second step is selecting the kind of plant as more interception and evapotranspiration as possible, last step is replacing impervious land with porous pavement.

• 한국경영과학회:학술대회논문집
• /
• 한국경영과학회 2001년도 추계학술대회 논문집
• /
• pp.150-153
• /
• 2001

The selection of the optimal partners in supply chain management is one of the most critical success factors. In the past, partners for outsourcing production were selected repeatedly within a closed group of candidates due to the limited information and location of partners. But, the wide use of internet and the development of electronic commerce make it possible that the partners capable of providing the optimal services are selected regardless of their location or nationality. And the concept of partners was limited to the provider of production resources. In the supply chain management, the concept should be extended to the provider of transportation and warehouse due to the high portion of transportation cost among the total production cost Therefore, In this study, we propose an analytical approach to the selection of Production and transportation partners in supply chain management. For this purpose a mathematical model is developed, and then a heuristic algorithm based on tabu search is presented since the model belongs to the NP hard problem.

• 한국통신학회논문지
• /
• 제36권1C호
• /
• pp.22-27
• /
• 2011

본 논문에서는 하향링크 다중 사용자 Multiple Input Multiple Output (MIMO)채널 환경에서 셀 용량을 증가시키기 위한 Zero Forcing 기반의 효율적 사용자 선택 기법을 제안한다. 이 기법에서는 채널 전력이 미리 결정된 한계점을 초과하는 사용자들을 주요 그룹으로 구성한다. 100명의 사용자와 4개의 기지국 안테나를 가정한 컴퓨터 모의실험을 통해서 제안된 기법이 기존 기법보다 0.33 bps/Hz 만큼 젤 용량을 추가로 제공할 수 있음이 밝혀졌다.

• 디지털산업정보학회논문지
• /
• 제14권3호
• /
• pp.77-85
• /
• 2018

Multi-User Multiple-Input Multiple-Output (MU-MIMO) is the core technology for improving the channel capacity compared to Single-User MIMO (SU-MIMO) by using multiuser gain and spatial diversity. Key problem for the MU-MIMO is the user selection which is the grouping the users optimally. To solve this problem, we adopt Extreme Value Theory (EVT) at the beginning of the proposed algorithm, which defines a primary user set instead of a single user that has maximum channel power according to a predetermined threshold. Each user in the primary set is then paired with all of the users in the system to define user groups. By comparing these user groups, the group that produces a maximum sum rate can be determined. Through computer simulations, we have found that the proposed method outperforms the conventional technique yielding a sum rate that is 0.81 bps/Hz higher when the transmit signal to noise ratio (SNR) is 30 dB and the total number of users is 100.

• 한국CDE학회논문집
• /
• 제3권3호
• /
• pp.177-182
• /
• 1998

The computer related techniques have been developed rapidly in the field of metal removal industry. The electronic catalog for mechanical elements is one of such developments realized in recent years. As the emergence of the electronic catalogs in the 90s, it breaks the restriction of the traditional catalog and becomes a helpful and efficient tool in the field. In this paper, we develop an electronic catalog for the cutting tools of the Korea Tungsten Co. Ltd. The catalog consists of three parts: a preprocessor for tool selection, a database for tool, and a postprocessor for search result. A preprocessor for tool selection is developed under a user-friendly consideration. A database for tool consists of milling, drilling, and reaming tools with a list of cutters, inserts, and components. A postprocessor for search result consists of weight, dimension. drawing, product order number etc. of the tools. It also suggests the optimized cutting condition of the selected tool using a neural network technique which is done by an independent research group.

• 산업공학
• /
• 제16권spc호
• /
• pp.70-75
• /
• 2003

There have been dramatic technology changes in telecommunication sector during last two decades. New application services like the introduction of international mobile services, fiber optical home services and satellite personal services are expected to play a major role in the whole field of telecommunication services. In the users' viewpoint, the proper selection and the efficient application of the specific telecommunication service among the various alternatives at hand are recognized as principle means for ensuring the whole business competitiveness. Despite the definite need for the telecommunication service selection strategies, very few researches have been reported in this field. In this paper, attempts are made to identify the companies' decision factors considered in selecting the telecommunication services, to draw the strategic dimensions and categorize each company with respect to the identified dimensions, and to finally find the characteristics of each strategic group in selecting the telecommunication service.

• 산업경영시스템학회지
• /
• 제23권54호
• /
• pp.13-26
• /
• 2000

Machining route selection to produce parts should be based on shop flow information because of input data at scheduling tasks and is one of the main problem in process planning. This paper addresses the problem of machining route selection in multi-stage process with machine group included a similar function. The model proposed is formulated as 0-1 integer programing considering the relation of parts and machine table size, avaliable time of each machine for planning period, and delivery date. The objective of the model is to minimize the sum of processing, transportation, and setup time for all parts. Genetic algorithm approach is developed to solve this model. The efficiency of the approach is examined in comparison with the method of branch and bound technique for the same problem. Also, this paper is to solve large problem scale and provide it if the multiple machining routes are existed an optimal solution.