• 한국멀티미디어학회논문지
• /
• 제13권12호
• /
• pp.1820-1831
• /
• 2010

When interacting with unknown environments, an autonomous agent needs to decide which action or action order can result in a good state and determine the transition probability based on the current state and the action taken. The traditional multiple sequential learning model requires predefined probability of the states' transition. This paper proposes a multiple sequential learning and prediction system with definition of autonomous states to enhance the automatic performance of existing AI algorithms. In sequence learning process, the sensed states are classified into several group by a set of proposed motivation filters to reduce the learning computation. In prediction process, the learning agent makes a decision based on the estimation of each state's cost to get a high payoff from the given environment. The proposed learning and prediction algorithms heightens the automatic planning of the autonomous agent for interacting with the dynamic unknown environment. This model was tested in a virtual library.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제8권1호
• /
• pp.123-144
• /
• 2014

Wireless sensor networks are generally deployed for specific applications to accomplish certain objectives over a period of time. To fulfill these objectives, it is crucial that the sensor network continues to function for a long time, even if some of its nodes become faulty. Energy efficiency and fault tolerance are undoubtedly the most crucial requirements for the design of an information extraction protocol for any sensor network application. However, most existing software agent based information extraction protocols are incapable of satisfying these requirements because of static agent itineraries and large agent sizes. This paper proposes an Information Extraction protocol based on Multiple software Agents with Dynamic Itineraries (IEMADI), where multiple software agents are dispatched in parallel to perform tasks based on the query assigned to them. IEMADI decides the itinerary for an agent dynamically at each hop using local information. Through mathematical analysis and simulation, we compare the performance of IEMADI with a well known static itinerary based protocol with respect to energy consumption and response time. The results show that IEMADI provides better performance than the static itinerary based protocols.

• Journal of Preventive Medicine and Public Health
• /
• 제46권5호
• /
• pp.213-225
• /
• 2013

Objectives: The aim of this study was to evaluate the association between Agent Orange exposure and self-reported diseases in Korean Vietnam veterans. Methods: A postal survey of 114 562 Vietnam veterans was conducted. The perceived exposure to Agent Orange was assessed by a 6-item questionnaire. Two proximity-based Agent Orange exposure indices were constructed using division/brigade-level and battalion/ company-level unit information. Adjusted odds ratios (ORs) for age and other confounders were calculated using a logistic regression model. Results: The prevalence of all self-reported diseases showed monotonically increasing trends as the levels of perceived self-reported exposure increased. The ORs for colon cancer (OR, 1.13), leukemia (OR, 1.56), hypertension (OR, 1.03), peripheral vasculopathy (OR, 1.07), enterocolitis (OR, 1.07), peripheral neuropathy (OR, 1.07), multiple nerve palsy (OR, 1.14), multiple sclerosis (OR, 1.24), skin diseases (OR, 1.05), psychotic diseases (OR, 1.07) and lipidemia (OR, 1.05) were significantly elevated for the high exposure group in the division/brigade-level proximity-based exposure analysis, compared to the low exposure group. The ORs for cerebral infarction (OR, 1.08), chronic bronchitis (OR, 1.05), multiple nerve palsy (OR, 1.07), multiple sclerosis (OR, 1.16), skin diseases (OR, 1.05), and lipidemia (OR, 1.05) were significantly elevated for the high exposure group in the battalion/company-level analysis. Conclusions: Korean Vietnam veterans with high exposure to Agent Orange experienced a higher prevalence of several self-reported chronic diseases compared to those with low exposure by proximity-based exposure assessment. The strong positive associations between perceived self-reported exposure and all self-reported diseases should be evaluated with discretion because the likelihood of reporting diseases was directly related to the perceived intensity of Agent Orange exposure.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제9권6호
• /
• pp.2050-2077
• /
• 2015

One of the major issues in emerging cloud management system needs the efficient service level agreement negotiation framework, with an optimal negotiation strategy. Most researchers focus mainly on the atomic service negotiation model, with the assistance of the Agent Controller in the broker part to reduce the total negotiation time, and communication overhead to some extent. This research focuses mainly on composite service negotiation, to further minimize both the total negotiation time and communication overhead through the pre-request optimization of broker strategy. The main objective of this research work is to introduce an Automated Dynamic Service Level Agreement Negotiation Framework (ADSLANF), which consists of an Intelligent Third-party Broker for composite service negotiation between the consumer and the service provider. A broker consists of an Intelligent Third-party Broker Agent, Agent Controller and Additional Agent Controller for managing and controlling its negotiation strategy. The Intelligent third-party broker agent manages the composite service by assigning its atomic services to multiple Agent Controllers. Using the Additional Agent Controllers, the Agent Controllers manage the concurrent negotiation with multiple service providers. In this process, the total negotiation time value is reduced partially. Further, the negotiation strategy is optimized in two stages, viz., Classified Similarity Matching (CSM) approach, and the Truncated Negotiation Group Gale Shapely Stable Matching (TNGGSSM) approach, to minimize the communication overhead.

• Journal of information and communication convergence engineering
• /
• 제20권3호
• /
• pp.181-188
• /
• 2022

An intelligent distributed multi-agent system (IDMS) using reinforcement learning (RL) is a challenging and intricate problem in which single or multiple agent(s) aim to achieve their specific goals (sub-goal and final goal), where they move their states in a complex and cluttered environment. The environment provided by the IDMS provides a cumulative optimal reward for each action based on the policy of the learning process. Most actions involve interacting with a given IDMS environment; therefore, it can provide the following elements: a starting agent state, multiple obstacles, agent goals, and a cluttered index. The reward in the environment is also reflected by RL-based agents, in which agents can move randomly or intelligently to reach their respective goals, to improve the agent learning performance. We extend different cases of intelligent multi-agent systems from our previous works: (a) a proposed environment-clutter-based-index for agent sub-goal selection and analysis of its effect, and (b) a newly proposed RL reward scheme based on the environmental clutter-index to identify and analyze the prerequisites and conditions for improving the overall system.

• 한국지능시스템학회논문지
• /
• 제10권6호
• /
• pp.584-590
• /
• 2000

Urban traffic situations are extremely complex and highly interactive. The multi-agent systems approach can provide a new desirable solution. Currently, a traffic simulator is needed to understand and explore the difficulties in an agent-oriented traffic control. This paper presents an agent-oriented fuzzy logic controller for multiple crossroads simulation. A fuzzy logic control simulation with variables of arrival, queue, and traffic volume could alleviate traffic congestion. We developed an agent-oriented simulator suitable for traffic junctions with η$\times$η intersections in Visual C++. The proposed method adaptively controls the cycle of traffic signals even though the traffic volume varies. The effectiveness of this method was shown through simulation of multiple intersections.

• 대한전자공학회:학술대회논문집
• /
• 대한전자공학회 2006년도 하계종합학술대회
• /
• pp.925-926
• /
• 2006

In the multi-agent system for a single task, the action selection can be made for the real-time environment by using the global coordination space, global coordination graph and MAX-PLUS algorithm. However, there are some difficulties in multi-agent system for multi-tasking. In this paper, a real-time decision making method is suggested by using coordination space, coordination graph and dynamic coordinated state of multi-agent system including many agents and multiple tasks. Specifically, we propose locally dynamic coordinated state to effectively use MAX-PLUS algorithm for multiple tasks completion. Our technique is shown to be valid in the box pushing simulation of a multi-agent system.

• 한국멀티미디어학회논문지
• /
• 제12권12호
• /
• pp.1819-1826
• /
• 2009

To enhance the automatic performance of existing predicting and planning algorithms that require a predefined probability of the states' transition, this paper proposes a multiple sequence generation system. When interacting with unknown environments, a virtual agent needs to decide which action or action order can result in a good state and determine the transition probability based on the current state and the action taken. We describe a sequential behavior generation method motivated from the change in the agent's state in order to help the virtual agent learn how to adapt to unknown environments. In a sequence learning process, the sensed states are grouped by a set of proposed motivation filters in order to reduce the learning computation of the large state space. In order to accomplish a goal with a high payoff, the learning agent makes a decision based on the observation of states' transitions. The proposed multiple sequence behaviors generation system increases the complexity and heightens the automatic planning of the virtual agent for interacting with the dynamic unknown environment. This model was tested in a virtual library to elucidate the process of the system.

• 한국정보통신학회논문지
• /
• 제5권1호
• /
• pp.173-182
• /
• 2001

본 논문에서는 이종 분산 지식기반 시스템을 구축하는데에 온톨로지가 어떤 역할을 하는지에 대해 논의한다. 먼저 다중 에이전트 시스템 기반 지식 공유와 재사용의 구조물인 지식사회에서 온톨로지와 에이전트간 관계를 논의하는데 온톨로지는 각 에이전트에 있어 지식사회에 연결되기 위한 최소한의 요구조건이다. 둘째로 지식사회에서 온톨로지가 어떻게 사용되고 있는가를 보여주기 위해 온톨로지에 의한 중재를 설명한다. 중재자라 불리는 특별한 에이전트는 주소가 없는 메시지들을 분석해서 온톨로지와 상의하고 온톨로지와 에이전트간 관계를 참고로 해서 수신후보 에이전트들을 추론한다. 셋째, 개념화 방식을 나타낼 수 있는 각 상황들의 결합으로서 온톨로지를 모델링한다. 상황통합을 의미하는 결합상황 혹은 상황선택을 의미하는 범주 상황으로서 상황들이 결합된다. 상황에 의한 온톨로지는 실세계 현상에 대한 이종 다중 기술을 허용하기 때문에 이종 지식기반시스템에게는 적합하다. 또한 다중상황을 해석하는 방식으로서 메시지 번역을 보여준다. 번역에이전트는 상황들의 종속성을 분석하여 어떤 상황을 가진 메시지를 다른 상황을 가진 메시지로 번역할 수 있다. 에이전트들을 쉽고 자연스럽게 구축하기 위해서 메시지의 중재와 번역은 중요한데 각 에이전트에 대하여 다른 에이전트들에 대한 지식이 덜 요청되기 때문이다.

• International Journal of Internet, Broadcasting and Communication
• /
• 제14권2호
• /
• pp.192-198
• /
• 2022

Companies are building and utilizing their own data analysis systems according to business characteristics in the distributed cloud. However, as businesses and data types become more complex and diverse, the demand for more efficient analytics has increased. In response to these demands, in this paper, we propose an unsupervised learning-based data analysis agent to which reinforcement learning is applied for effective data analysis. The proposal agent consists of reinforcement learning processing manager and unsupervised learning manager modules. These two modules configure an agent with k-means clustering on multiple nodes and then perform distributed training on multiple data sets. This enables data analysis in a relatively short time compared to conventional systems that perform analysis of large-scale data in one batch.