• Proceedings of the Korean Information Science Society Conference
• /
• 2007.06c
• /
• pp.319-323
• /
• 2007

실세계의 여러 문제들은 마르코프 결정 문제(Markov decision problem, MDP)로 표현될 수 있고, 이 MDP는 모델이 알려진 경우에는 평가치 반복(value iteration) 이나 모델이 알려지지 않은 경우에도 강화 학습(reinforcement learning) 알고리즘 등을 사용하여 풀 수 있다. 하지만 이들 알고리즘들은 시간 복잡도가 높아 크기가 큰 실세계 문제에 적용하기 쉽지 않아, MDP를 계층적으로 분할하거나, 여러 단계를 묶어서 수행하는 등의 시간적 추상화(temporal abstraction) 방법이 제안되어 왔다. 이러한 시간적 추상화 방법들의 문제점으로는 시간적 추상화의 디자인에 따라 MDP의 풀이 성능이 크게 달라질 수 있으며, 많은 경우 사용자가 이 디자인을 직접 제공해야 한다는 것들이 있다. 최근 사용자의 간섭이 필요 없이 자동적으로 시간적 추상화를 만드는 방법들이 제안된 바 있으나, 이들 방법들 역시 결과물에 대한 이론적인 성능 보장(performance guarantee)은 제공하지 못하고 있다. 본 연구에서는 이러한 문제점을 해결하기 위해 MDP의 구조와 그 풀이 성능을 연관짓는 복잡도 척도에 대해 살펴본다. 이를 위해 MDP로부터 얻은 상태 경로 그래프(state trajectory graph)의 위상적 성질들을 여러 네트워크 척도(network measurements) 들을 이용하여 측정하고, 이와 MDP의 풀이 성능과의 관계를 다양한 상황에 대해 실험적, 이론적으로 분석해 보았다.

• Annual Conference of KIPS
• /
• 2014.11a
• /
• pp.1126-1129
• /
• 2014

사회관계망 분석에 있어서 매개중심도는 네트워크를 구성하는 노드들의 상대적인 중요도를 파악하기 위한 척도로서 오랫동안 사용되어 왔다. 그러나 이러한 매개중심도를 계산하기 위한 계산 복잡도가 높기 때문에 대규모 사회관계망에서는 매개중심도를 계산하기란 쉽지 않은 문제이다. 본 논문에서는 네트워크를 구성하는 각각의 노드들마다 자신의 지역정보를 활용하여 구성한 네트워크에서 매개중심도를 산출함으로써 시간복잡도를 줄이는 한편 지역정보 기반의 네트워크의 특징을 분석함으로써 매개중심도를 더 빠르게 산출할 수 있는 알고리즘을 제안한다. 그리고 실제 소셜 네트워크에서의 실험을 통하여 제안 알고리즘이 기존 알고리즘에 비해 매개중심도를 더 빠르게 산출함을 보인다.

• Journal of Korean Society of Transportation
• /
• v.26 no.2
• /
• pp.147-156
• /
• 2008

It has been widely known that there are so many factors making travel demand errors in transportation forecasting steps. One of the reasons may stem from the level of aggregation of zone and network in analysis process. This paper investigates the effect of level of aggregation considering with number of zones in travel demand forecasting by expanding or reducing the zone and network gradually. Numerical results show that the aggregation could not make a significant impact on the travel demand, while disaggregation does. These results imply that a careful manipulation is required to add or to reduce zones and links in transportation planning process.

• The Journal of the Korea Contents Association
• /
• v.16 no.9
• /
• pp.127-135
• /
• 2016

We identified a competition network which has never been studied before and investigated the relationship between centrality of participants in singing competition and their winning percentage within the competition network. We collected competition data from 'Immortal Songs: Singing the Legend', which is a Korean television music competition program, and constructed a competition network. We calculated centrality and winning percentage and analyzed their relationship using correlation analysis, regression analysis, and visualization. There are four main findings in this research. First, a competition network is a scale-free network whose degree distribution follows a power law. Second, there is a logarithmic relationship between the count of competition and closeness. Third, winning percentage converges to approximately 60% for players who have participated in more than 20 competitions. Lastly, a strength of opponents affects approximately 23% of winning percentage for players with less than 20 competitions. The academic significance of this study is that we pioneered the definition of the competition network and applied social network analysis method. Another significant contribution of this paper is that we found explicit patterns between the centrality and winning percentage, suggesting ways to improve social relationship in competition network and to increase winning percentage.

• Journal of the Korean Society for Library and Information Science
• /
• v.40 no.2
• /
• pp.333-355
• /
• 2006

Network generation methods to visualize bibliometric data for examining the intellectual structure of knowledge domains are investigated in some detail. Among the four methods investigated in this study, pathfinder network algorithm is the most effective method in representing local details as well as global intellectual structure. The nearest neighbor graph, although never used in bibliometic analysis, also has some advantages such as its simplicity and clustering ability. The effect of input data preparation process on resulting intellectual structures are examined, and concluded that unlike MDS map with clusters, the network structure could be changed significantly by the differences in data matrix preparation process. The network generation methods investigated in this paper could be alternatives to conventional multivariate analysis methods and could facilitate our research on examining intellectual structure of knowledge domains.

• Journal of the Korea Society of Computer and Information
• /
• v.27 no.2
• /
• pp.33-42
• /
• 2022

In this paper, we proposed a deep learning model based on multi-scale dilated convolution feature fusion for the segmentation of black ice region in road image to send black ice warning to drivers in real time. In the proposed multi-scale dilated convolution feature fusion network, different dilated ratio convolutions are connected in parallel in the encoder blocks, and different dilated ratios are used in different resolution feature maps, and multi-layer feature information are fused together. The multi-scale dilated convolution feature fusion improves the performance by diversifying and expending the receptive field of the network and by preserving detailed space information and enhancing the effectiveness of diated convolutions. The performance of the proposed network model was gradually improved with the increase of the number of dilated convolution branch. The mIoU value of the proposed method is 96.46%, which was higher than the existing networks such as U-Net, FCN, PSPNet, ENet, LinkNet. The parameter was 1,858K, which was 6 times smaller than the existing LinkNet model. From the experimental results of Jetson Nano, the FPS of the proposed method was 3.63, which can realize segmentation of black ice field in real time.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.34 no.11B
• /
• pp.1216-1224
• /
• 2009

This paper presents cross-layer based packet scheduling and routing algorithm to effectively transmit delay-sensitive multimedia data over mobile ad hoc network. At first, packet urgency, node urgency, and route urgency are defined based on the end-to-end delay requirement of each packet and the number of hops over a route. Based on the urgency metric, the proposed transmission scheme is that packet scheduling algorithm at the MAC layer and routing algorithm at the network layer are tightly-coupled to improve the transmission performance for delay-sensitive multimedia. Finally, simulation results are provided to show the performance of proposed transmission scheme.

• Journal of the Korea Society for Simulation
• /
• v.20 no.2
• /
• pp.1-10
• /
• 2011

In queueing network analysis, arrival processes are usually modeled as renewal processes by matching mean and variance. The renewal approximation simplifies the analysis and provides reasonably good estimate for the performance measures of the queueing systems under moderate conditions. However, high variability in arrival process or in service process requires more sophisticated approximation procedures for the variability parameter of departure/arrival processes. In this paper, we propose an heuristic approach to refine Whitt's variability function with the k-interval squared coefficient of variation also known as the index of dispersion for intervals(IDI). Regression analysis is used to establish an empirical relationships between the IDI of arrival process and the IDI of departure process of a queueing system.

• The Journal of The Korea Institute of Intelligent Transport Systems
• /
• v.11 no.6
• /
• pp.15-22
• /
• 2012

A performance index of singalized intersections is a standard to optimize signal control variables and to manage traffic flow. Traffic delays is generally used to minimize the average delay time on intersections or networks, progression efficiency is used to improve travel speed of main cooridors or to provide transit signal priority. We manage traffic flows with only selecting one index between delays and progression according to the objective of traffic management and field characteristics. In real field, the driver's satisfaction is high in any performance criteria when the waiting time is shorter and the unnecessary stop in front of traffic is smaller. This paper aims to develop simulation model to represent real progression with concurrently considering delays and progression. In order to reflect an effect of level of traffic volumes and residual queues which don't be considered in prior progression model, we apply shockwave model with flow-density diagram. We derive Cell Transmission Model of Daganzo in order to develop the delay index and the progression index for the macroscopic simulation model. In order to validate the effect, we analysis traffic delays and progression efficiency with comparing this model to Transyt-7F and PASSER V.

• Journal of Korea Society of Industrial Information Systems
• /
• v.12 no.5
• /
• pp.1-13
• /
• 2007

In this research, we found out that bridging nodes have great effect on the robustness of protein-protein interaction networks. Until now, many researchers have focused on node's degree as node's essentiality. Hub nodes in the scale-free network are very essential in the network robustness. Some researchers have tried to relate node's essentiality with node's betweenness centrality. These approaches with betweenness centrality are reasonable but there is a positive relation between node's degree and betweenness centrality value. So, there are no differences between two approaches. We first define a bridging node as the node with low connectivity and high betweenness value, we then verify that such a bridging node is a primary factor in the network robustness. For a biological network database from Internet, we demonstrate that the removal of bridging nodes defragment an entire network severally and the importance of the bridging nodes in the network robustness.