• 한국컴퓨터정보학회논문지
• /
• 제21권5호
• /
• pp.79-89
• /
• 2016

In this paper, we proposed the object classification method using genetic and dynamic random forest consisting of optimal combination of unit tree. The random forest can ensure good generalization performance in combination of large amount of trees by assigning the randomization to the training samples and feature selection, etc. allocated to the decision tree as an ensemble classification model which combines with the unit decision tree based on the bagging. However, the random forest is composed of unit trees randomly, so it can show the excellent classification performance only when the sufficient amounts of trees are combined. There is no quantitative measurement method for the number of trees, and there is no choice but to repeat random tree structure continuously. The proposed algorithm is composed of random forest with a combination of optimal tree while maintaining the generalization performance of random forest. To achieve this, the problem of improving the classification performance was assigned to the optimization problem which found the optimal tree combination. For this end, the genetic algorithm methodology was applied. As a result of experiment, we had found out that the proposed algorithm could improve about 3~5% of classification performance in specific cases like common database and self infrared database compare with the existing random forest. In addition, we had shown that the optimal tree combination was decided at 55~60% level from the maximum trees.

• 대한전자공학회:학술대회논문집
• /
• 대한전자공학회 1999년도 추계종합학술대회 논문집
• /
• pp.943-946
• /
• 1999

In this paper, we present a new clock routing algorithm which minimizes total wirelength under any given path-length skew bound. The algorithm onstructs a bounded-skew tree(BST) in two steps:(ⅰ) a bottom-up phase to construct a binary tree of shortest-distance feasible regions which represent the loci of possible placements of clock entry points, and (ⅱ) a top-down phase to determine the exact locations of clock entry points. Experimental results show that our clock routing algorithm, named BST/DME, can produce a set of solutions with skew and wirelength trade-off.

• 산업경영시스템학회지
• /
• 제23권61호
• /
• pp.33-39
• /
• 2000

In this paper, we present an algorithm for the multicast routing problem when there exit the cell replication limits of ATM switching nodes. This problem can be formulated as a Degree Constrained Minimum Steiner Tree Problem(DCSP). The proposed algorithm is a modification of the shortest path heuristic originally devised for minimum Steiner tree problem. From the experimental results, it can be seen that our algorithm is efficient to obtain a near optimal solution with comparatively low computational time.

• 대한전자공학회논문지
• /
• 제26권9호
• /
• pp.1340-1347
• /
• 1989

For the efficient recognition of Korean Alphabets, a tree structure discrimination algorithm employing NRF-SDF concept is proposed. This algorithm consists of several main-steps, which contain several sub-steps. Each step contains vowels or consonants for training image. This algorithm reduces processing and recognition time than any other conventional algorithms for recognition of Korean Alphabets. A simulation results indicated that this algorithm has a satisfactory performance.

• 한국경영과학회지
• /
• 제33권3호
• /
• pp.17-28
• /
• 2008

The Steiner arborescence problem is known to be NP-hard. The objective of this problem is to find a minimal Steiner tree which starts from a designated node and spans all given terminal nodes. This paper proposes a method based on a two-step procedure to solve this problem efficiently. In the first step, graph reduction rules eliminate useless nodes and arcs which do not contribute to make an optimal solution. In the second step. ant colony algorithm with use of Prim's algorithm is used to solve the Steiner arborescence problem in the reduced graph. The proposed method based on a two-step procedure is tested in the five test problems. The results show that this method finds the optimal solutions to the tested problems within 50 seconds. The algorithm can be applied to undirected Steiner tree problems with minor changes. 18 problems taken from Beasley are used to compare the performances of the proposed algorithm and Singh et al.'s algorithm. The results show that the proposed algorithm generates better solutions than the algorithm compared.

• 산업공학
• /
• 제22권3호
• /
• pp.278-286
• /
• 2009

The anti-collision algorithms to identify a number of tags in real-time in RFID systems are divided into the anti-collision algorithms based on the Framed slotted ALOHA that randomly select multiple slots to identify the tags, and the anti-collision algorithms based on the Tree-based algorithm that repeat the questions and answer process to identify the tags. In the hybrid algorithm which is combined the advantages of these algorithms, tags are distributed over the frames by selecting one frame among them and then identified by using the Query tree frame by frame. In this hybrid algorithm, however, the time of identifying all tags may increase if many tags are concentrated in a few frames. In this study, to improve the performance of the hybrid algorithm, we suggest an improved algorithm that the tags select a specific group of frames based on the earlier bits of the tag ID so that the tags are distribute equally over the frames. By using the simulation and mathematical analysis, we show that the suggested algorithm outperforms traditional hybrid algorithm from the viewpoint of the number of queries per frame and the time of identifying all tags.

• 인터넷정보학회논문지
• /
• 제21권4호
• /
• pp.35-41
• /
• 2020

본 논문은 Context Tree Weighting (CTW) 를 이용하여 Adaptive Multi-Rate (AMR) 데이터의 압축 성능을 개선하는 알고리즘을 제안한다. AMR은 IMT-2000에서 채택된 음성부호화 표준안으로써, 무선채널의 환경변화에 대처할 수 있도록 4.75 kbit/s 에서 12.2 kbit/s 까지 8가지의 전송률을 지원한다. CTW는 산술부호화기의 일종으로, 가변 차수 마르코프 모델을 사용하는 압축기이다. 우리는 CTW가 비트단위로 수행한다는 점을 고려하여 AMR 데이터를 변환한 후 CTW로 압축하는 알고리즘을 제안한다. 제안하는 알고리즘의 유효성을 검증하기 위하여 ZIP을 포함한 기존 압축방식과 제안된 알고리즘의 압축률을 비교하는 실험을 하였다. 실험 결과, AMR 데이터의 평균 추가 압축률이 ZIP의 경우 약 3.21%, 제안된 알고리즘의 경우 약 9.10%로 나타났다. 따라서 본 논문에서 제안한 알고리즘이 AMR 데이터의 압축 성능을 약 5.89% 개선하였다.

• 한국산학기술학회논문지
• /
• 제21권10호
• /
• pp.77-83
• /
• 2020

국토를 효율적으로 관리하고 도시문제를 과학적으로 해결하기 위해 최근 스마트시티, 디지털트윈 등 3차원 공간정보 관련 기술이 급격하게 발전하고 있다. 이러한 3차원 공간정보 구축은 주로 영상정보를 이용하여 객체를 3차원 입체화하고 실감형 영상인 텍스처링 영상을 추출하여 객체벽면에 영상을 부여하는 방식으로 수행된다. 하지만 객체 주변의 다양한 요인으로 인해 텍스처링 영상에서는 필연적으로 폐색영역이 발생한다. 이에 본 연구에서는 최근 기술인 딥러닝 기술 중에서 ResNet 알고리즘을 이용하여 건물 폐색을 유발하는 가로수에 대한 데이터셋을 만들고 이에 대한 해결방안을 제시하고자 한다. 연구결과 ResNet 알고리즘의 공간정보 적용 가능성을 판단하고 이를 적용한 레이블링 생성 SW 개발하여 실제 가로수를 대상으로 데이터셋을 구축하였다. 구축된 데이터셋을 텍스처링 영상에 적용하여 정확도와 재현율로 검출능력을 분석하였다. 분석결과를 위해 딥러닝 분야에서 많이 사용되고 있는 정밀도와 재현율을 이용한 F값을 적용하였으며 가로수 단일 객체가 포함된 건물의 측면부 영상과 경사 영상에 대해서는 높은 F값을 도출하여 우수한 성과를 확인하였으나, 같은 해상도를 가진 건물 전면부 영상에서는 그림자 등의 요인으로 F값이 낮음을 확인하였다.

• ETRI Journal
• /
• 제28권3호
• /
• pp.329-336
• /
• 2006

This paper investigates the problem of protecting multicast sessions in mesh wavelength-division multiplexing (WDM) networks against single link failures, for example, a fiber cut in optical networks. First, we study the two characteristics of multicast sessions in mesh WDM networks with sparse light splitter configuration. Traditionally, a multicast tree does not contain any circles, and the first characteristic is that a multicast tree has better performance if it contains some circles. Note that a multicast tree has several branches. If a path is added between the leave nodes on different branches, the segment between them on the multicast tree is protected. Based the two characteristics, the survivable multicast sessions routing problem is formulated into an Integer Linear Programming (ILP). Then, a heuristic algorithm, named the adaptive shared segment protection (ASSP) algorithm, is proposed for multicast sessions. The ASSP algorithm need not previously identify the segments for a multicast tree. The segments are determined during the algorithm process. Comparisons are made between the ASSP and two other reported schemes, link disjoint trees (LDT) and shared disjoint paths (SDP), in terms of blocking probability and resource cost on CERNET and USNET topologies. Simulations show that the ASSP algorithm has better performance than other existing schemes.

• 한국컴퓨터정보학회논문지
• /
• 제19권7호
• /
• pp.131-140
• /
• 2014

본 논문에서는 최소신장트리를 구하는 크루스칼 알고리즘의 효율적인 구현 방법을 제시한다. 제시하는 방법은 union-find 자료구조를 이용하며, 노드 집합을 나타내는 각 트리의 깊이를 줄이기 위해 union 연산시 루트까지의 경로에 있는 노드들의 위치를 최종 루트의 자식노드로 직접 이동하여 깊이를 줄이도록 하는 방법이다. 이 방법은 루트까지의 경로를 축소하고 노드의 레벨을 축소시킴으로써 트리의 깊이도 줄일 수 있다. 트리의 깊이가 줄어든다면 노드가 속하는 트리의 루트를 찾는 시간을 줄일 수 있게 되어 효율적인 방법이라 할 수 있다. 본 장에서 제안하는 방법을 그래프로 평가해보고 분석해 본 결과, 기존의 union() 방법이나 경로축소방법인 union2() 보다 트리의 깊이를 작게 유지함을 알 수 있다.