• The KIPS Transactions:PartD
• /
• v.8D no.4
• /
• pp.439-450
• /
• 2001

Mobile agent is one of the promising ways of overcoming network load and latency. It is also a new software paradigm including those concepts of client/server, applet/servelet, and code-on-demand. In this paper, a new mobile agent concept with flexible replay scheme is proposed, which can deal with embarrassing situation when replying results should be accomplished in various distributed applications. For example, they are 1) a burden of bulky result, 2) a need of different replay scheme dependent on work type, 3) connection failure to the target node, and 4) a need of fast migration to next node to keep the itinerary. Regarding the source node, there may be another situation that it wants to be aware of whether its wi가 is completed or not. Three kinds of reply schemes are proposed for dealing flexibly with such situations. They are based on priority reply list where nodes to be replied are stored according to their priorities. An experimental supply chain management model using the proposed reply schemes is designed and implemented. It is one of good distributed applications appropriate for our reply schemes, because it requires different reply schemes according to work types and it is important to gather required information in time.

• Proceedings of the IEEK Conference
• /
• 2003.07a
• /
• pp.382-385
• /
• 2003

Ad hoc 네트워크는 고정된 기반구조 없이 이동 노드들로만 구성된 네트워크를 가리킨다. 노드들의 이동으로 인해 네트워크 토폴로지는 예측할 수 없게 자주 변할 수 있다. 그 결과로 기 설정되어 있는 경로의 단절이 발생할 수 있으며, 새로운 경로가 설정될 때까지 데이터가 손실될 수 있다. 존재하는 경로의 단절로 인해 손실되는 패킷의 수를 줄이기 위해, 본 논문은 multiple-reply Ad hoc On-demand Distance Vector (mrAODV) 방식을 제안한다. 본 논문의 알고리즘은 발신지와 수신지 사이에 다중 경로를 설정한다. ns-2 시뮬레이션을 사용하여 제안된 방식에 대한 성능을 평가한다.

• 한국HCI학회:학술대회논문집
• /
• 2006.02a
• /
• pp.1298-1304
• /
• 2006

공급체인관리(Supply Chain Management)는 단순히 물질의 흐름관리뿐만 아니라 제품의 생산, 공급, 유통 등 다양하고 복잡한 활동들의 물질 및 정보 흐름의 통합이라고 할 수 있다. 그러므로 각 구성원간의 긴밀한 협력과 공동작업을 위해서는 원활한 정보 수집 및 응답이 요구된다. 하지만 기존 시스템은 비이동성 에이전트 개념으로 네트워크의 오버헤드나 혹은 접속 단절 상황에서 작업 수행이 불가능하여 구성원간의 정보 수집과 응답이 실제 업무 처리에 지연이 발생되므로 생산에서 판매까지의 리드타임이 길어지게 되고 채찍효과(Bullwhip Effect)가 더욱 크게 나타난다. 따라서 본 논문에서는 구성원간 협력과 공동작업을 위해 효율적으로 처리되어야 할 정보 수집 및 응답을 대상으로 하는 효율적 다중 응답을 위한 이동 에이전트 시스템을 설계한다. 제안시스템은 기존 시스템의 문제 해결을 위해 이동 에이전트와 객체 복제 메커니즘을 제시하며, 또한 각 구성원 시스템의 투명성을 위해 네이밍 서비스의 구성 방법과 등록된 구성원 시스템 객체의 객체 참조자 정보에 따라 복제된 이동 에이전트가 각 구성원 시스템으로 이주하는 이주 방식과 이주 후, 개별적인 작업 수행 결과를 해당 각 구성원 시스템에 제공하는 다중 응답 시나리오를 제시한다.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.17 no.11
• /
• pp.2569-2574
• /
• 2013

In 13.56MHz RFID system of Auto-ID center, STAC protocol is defined as an anti-collision algorithm for multiple tag reading. In STAC protocol, the reader uses the tag number estimation scheme to determine the number of slots for the next reply round. If errors occur in the estimated number of tags, the identification performances will be degraded due to a lot of collision or idle slots. Therefore, in this paper, we propose an ERPB algorithm that the reader cancels the reply round when it experiences a series of collision or idle slots during the current round. The transmission probability control for the tag is also applied to the proposed algorithm. Through simulations, it is demonstrated that the collision rate for the proposed scheme is about 39% lower than STAC protocol. Therefore, the proposed scheme can achieve faster tag identification time compared with STAC protocol due to the low collision rate.

• 제어로봇시스템학회:학술대회논문집
• /
• 2005.06a
• /
• pp.1299-1303
• /
• 2005

Over the last decade the research of end-to-end behavior on computer network has grown by orders but it has few researching in hop-by-hop behavior. We think if we know hop-by-hop behavior it can make better understanding in network behavior. This paper represent ICMP time stamp request and time stamp reply as tool of network study for learning in hop-by-hop behavior to estimate link bandwidth and link structure. We describe our idea, experiment tools, experiment environment, result and analysis, and our discussion in our observative.

• Proceedings of the PSK Conference
• /
• 2003.10b
• /
• pp.68.2-68.2
• /
• 2003

We performed this survey to hear Yeungnam University Graduates' opinion on the current curriculum for the purpose of creating an education system of “practically competent person” instead of simply “competent person”. Questionnaire was made up of 13 multiple-choice questions and 1 descriptive question by Curriculum Reform Task Force of Yeungnam University. The survey was administered to randomly chosen 50 graduates by e-mail on August 1, 2003 and was collected between August 1 and August 10, 2003 for the analysis of respondent’s reply. (omitted)

• Journal of Korea Multimedia Society
• /
• v.16 no.1
• /
• pp.75-86
• /
• 2013

This paper presents a technique to implement embedded systems using interrupts of the one-chip microcontroller with many peripherals based on a multiple finite state machines model. The multiple finite state machine model utilizes the structure of FSMD used for hardware design and the features of flow control by interrupts. The main finite state machine corresponds to the main program and the sub-state machines corresponds to the interrupt subroutines. Therefore, interrupts from the peripherals can be processed immediately in the sub-state machines. The request and reply variables are used to interface between the finite state machines. Additional operating system is not necessary for the context switching between the main state machine and the sub-state machine, because the flow-control caused by interrupt can be replaced with the switching. An embedded system modeled on multiple finite state machine with ASM charts can be easily implemented by the conversion of ASM charts into C-language programs. This implementation technique can be easily adopted to the hardware oriented embedded systems because of the detail description of the model and the fast response to the interrupt events in the sub-state machine.

• The Research Journal of the Costume Culture
• /
• v.14 no.5
• /
• pp.850-863
• /
• 2006

The purpose of this study was to investigate if star marketing of on-line shopping malls affects consumers' WOM effect, and to compare the differences of consumption behavior between female teenagers and college students. Two hundred five female teenagers and college students who had purchased fashion goods through internet shopping mall participated in this study. For data analysis, descriptive statistics, factor analysis, t-test, and multiple regression were used. As the results, first, recognition of celebrity advertising models was classified into three factors such as 'trust of product', 'attractiveness of product' and 'leading interest of product' factors. Second, the greater exposure to celebrity models, the greater the good feelings about them, showing respondents' positive consumption behavior. Third, results of multiple regression revealed that behavior of pursuing celebrities' style accounted for 37% of the explained variance WOM behavior. Finally, t-test revealed that female college students were affected more by celebrity style and bought fashion items than female teenagers. However, female teenagers conducted more WOM behavior than college students. Based on these results, on-line fashion marketers would use these data for more their efficient fashion marketing strategies.

• Journal of information and communication convergence engineering
• /
• v.8 no.3
• /
• pp.307-311
• /
• 2010

EPCglobal Class-1 Gen-2 protocol has been proposed for UHF-band RFID systems. In Gen-2 standard, Q-algorithm was proposed to select a frame size for the next query round without estimating the number of tags. Therefore, the Q-algorithm has advantage that the reader's algorithm is simpler than other algorithms. However, it is impossible to allocate the optimized frame size. Also, the original Q-algorithm did not define an optimized parameter C for adjusting the frame size. In this paper, we propose an adaptive Q-algorithm with the different parameter $C_c$ and $C_i$ in accordance with the status of reply slot. Simulation results show that the proposed adaptive Q-algorithm outperforms the original Gen-2 Q-algorithm.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.34 no.12B
• /
• pp.1444-1451
• /
• 2009

This paper first proposes diverse C determining models in the Q algorithm which is proposed in the EPCglobal C1 Gen 2 standard and then compares and analyzes its performance. EPCglobal C1 Gen 2 standard proposes the slot-count (Q) selection algorithm for multiple tag identification environment, but there is no such definition for the C value which modifies the Q value depending on collision or no reply. During the tag anti-collision process, the Q algorithm adds C to the Q when there is a collision and reduces the Q by C when there is no reply. The modified Q value updates new slot-counts for tags which determines the tag identification speed, so the C value is an important factor. However, many researches only intend to increase the tag identification speed by proposing a new method or modifying the Q algorithm without any research about the C value. This paper suggests diverse C models which satisfies the EPCglobal C1 Gen 2 and analyzes their performance in the multi tag identification environment. The result of this paper can be used as an index for future researches on EPCglobal C1 Gen 2 C models and multiple tag identification performance.