• 제어로봇시스템학회:학술대회논문집
• /
• 2001.10a
• /
• pp.164.1-164
• /
• 2001

Q-learning is a recent reinforcement learning algorithm that does not need a modeling of environment and it is a suitable approach to learn behaviors for autonomous agents. But when it is applied to multi-agent learning with many I/O states, it is usually too complex and slow. To overcome this problem in the multi-agent learning system, we propose the successive Q-learning algorithm. Successive Q-learning algorithm divides state-action pairs, which agents can have, into several Q-functions, so it can reduce complexity and calculation amounts. This algorithm is suitable for multi-agent learning in a dynamically changing environment. The proposed successive Q-learning algorithm is applied to the prey-predator problem with the one-prey and two-predators, and its effectiveness is verified from the efficient avoidance ability of the prey agent.

• Journal of information and communication convergence engineering
• /
• v.17 no.3
• /
• pp.167-173
• /
• 2019

This study investigates the problem of leader-following consensus for multi-agent systems with output saturations. This study assumes that the agents are described as a neutrally stable system, and the leader agent generates the bounded trajectory within the saturation level. Then, the objective of the leader-following consensus is to track the trajectory of the leader by exchanging information with neighbors. To solve this problem, we propose an observer-based distributed consensus algorithm. Then, we provide a consensus analysis by applying the Lyapunov stability theorem and LaSalle's invariance principle. The result shows that the agents achieve the leader-following consensus in a global sense. Moreover, we can achieve the consensus by choosing any positive control gain. Finally, we perform a numerical simulation to demonstrate the validity of the proposed algorithm.

• Transactions of the Korean Society of Automotive Engineers
• /
• v.12 no.5
• /
• pp.125-135
• /
• 2004

The validity of simulation has been well-established for decades in areas such as computer and communication system. Recently, the technique has become entrenched in specific areas such as transportation and traffic forecasting. Several methods have been proposed for investigating complex traffic flows. However, the dynamics of vehicles and their driver's characteristics, even though it is known that they are important factors for any traffic flow analysis, have never been considered sufficiently. In this paper, the traffic simulation using a multi-agent approach with considering vehicle dynamics is proposed. The multi-agent system is constructed with the traffic environment and the agents of vehicle and driver. The traffic environment consists of multi-lane roads, nodes, virtual lanes, and signals. To ensure the fast calculation, the agents are performed on the based of the rules to regulate their behaviors. The communication frameworks are proposed for the agents to share the information of vehicles' velocity and position. The model of a driver agent which controls a vehicle agent is described in the companion paper. The vehicle model contains the nonlinear subcomponents of engine, torque converter, automatic transmission, and wheels. The simulation has proceeded for an interrupted and uninterrupted flow model. The result has shown that the driver agent performs human-like behavior ranging from slow and careful to fast and aggressive driving behavior, and that the change of the traffic state is closely related with the distance and the signal delay between intersections. The system developed shows the effectiveness and the practical usefulness of the traffic simulation.

• Journal of Intelligence and Information Systems
• /
• v.11 no.3
• /
• pp.129-139
• /
• 2005

The rapid growth of IT technologies enables the quality of human's daily life to be improved dramatically, Contrasting to the services in previous computing environment directed by user's request, the services in ubiquitous computing era of new IT technology are provided through recognizing users intention and utilizing context-aware information suited to the user. According to the contextual information, agents need to set a dynamic goal to achieve and work collaboratively with other agents. Agents that take control over their behaviors with capability of communicating with other agents become a thrust in this up-coming computing environment. This work focuses on building ontologies, shared knowledge bases among agents, to improve semantical interoperability among agents. More attention is given to the construction and effective management of ontology repository along with its requirement and organization. Ontology agent suggested takes an initiative role to manage the repository in a distributed manner and to facilitate the use of ontology in a multi-agent environment.

• Journal of Korea Game Society
• /
• v.9 no.5
• /
• pp.117-125
• /
• 2009

In real-time strategy simulation (RTS) game each team is composed of agents and executes strategies to win other team. Strategy needs agents' cooperation in each team. This needs multi-agent system (MAS). Centralized decision making, one of decision making in MAS, selects actions not to agents but to team by a coordinated agent. Decentralized decision making costs high because each agent communicates with each other. In this paper we propose a system which controls agents by grouping and allocates roles through negotiation by a coordinated agent. Then, when one of allocated actions is not executed or failed, a coordinated agent allocates its role to another agent. We make experiments in starcraft, famous RTS game. When a proposed method is applied, the performance of attack and defense is increased. The improved agents' team wins eight times per ten games.

• The Transactions of the Korea Information Processing Society
• /
• v.6 no.12
• /
• pp.3548-3558
• /
• 1999

There are many kinds of multi-agent frameworks which support the high-level knowledge representation languages for providing intelligence to their agents. But, the agent programming interfaces of the frameworks require to use some general-purpose programming languages as well as tile knowledge representation languages. In general, knowledge representation languages and general-purpose programming languages are different in their levels and data representation models. The differences can make the problems about tile coupling of the elements which are necessary for developing intelligent agents. This paper describes a new type of intelligent agent shell INAS(INtelligent Agent Shell) version 2 which has developed to cope with the problems. Unlike the previous agent frameworks, INAS supports a high-level integrated agent building language for building intelligent agents by itself. Therefore, the development of intelligent agents by using INAS version 2 does not suffer from the problems of the previous agent frameworks. Through the development of several intelligent agents, we experienced that the agent building language of INAS version 2 could reduce the difficulties of developing intelligent agents.

• Restorative Dentistry and Endodontics
• /
• v.23 no.1
• /
• pp.413-423
• /
• 1998

Dental ceramics exhibit excellent esthetic property, compressive strength, chemical durability, biocompatibility and translucency. This study evaluated the shear bond strength of composite resin to the new heat-pressed ceramic material (IPS-Empress System) depending on the surface treatments and bonding agents. The surface treatments were etching with 4.0% hydrofluoric acid, application of silane, and the combination of the two methods. Composite resin was bonded to ceramic with four kinds of dentin bonding agents(All-Bond 2, Heliobond, Scotch bond Multi-purpose and Tenure bonding agents). The ceramic specimen bonded with composite resin was mounted in the testing jig, and the universal testing machine(Zwick 020, Germany) was used to measure the shear bond strength with the cross head speed of 0.5 mm/min. The results obtained were as follows 1. The mean shear bond strength of the specimens of which the ceramic surface was treated with the combination of hydrofluoric acid and silane before bonding composite resin was significantly higher than those of the other surface treatment groups(p<0.05). 2. In the case of All-Bond 2 and Scotchbond Multi-purpose bonding agent group, the surface treatment methods did not influenced significantly on the shear bond(p>0.05). 3. Of the four bonding agents tested, the shear bond strength of Heliobond was significantly lower than those of other bonding agents regardless of the surface treatment methods(p<0.05). 4. The highest shear bond strength($12.55{\pm}1.92$ MPa) was obtained with Scotchbond Multipurpose preceded by the ceramic surface treatment with the combination of 4% hydrofluoric acid and silane.

• Asia pacific journal of information systems
• /
• v.10 no.2
• /
• pp.23-49
• /
• 2000

We proposed a conceptual design of the web-based agent model for global supply chain management(GSCM), where agents representing autonomous operational units, such as suppliers, factories, distribution center and customers, cooperate and are coordinated through the information exchange. The agent model assumed the hierarchical federated system. In the federated system, the agents of the same region are grouped and linked to the region-specific facilitator only through which communication between agents is allowed. The facilitator is responsible for monitoring and controlling the conversations consisting of the message flows across the agents. A web-based user presentation was also designed so that human users could involve in collaborative settings into the GSCM multi-agent system. In the conversation protocols which allow for complex coordinated behavior among agents, the KQML was extended to represent the messages. A GSCM scenario where the supply chain is formed upon customer order and supply decision is made was used to demonstrate the dynamics of the conversation protocols.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.25 no.1
• /
• pp.102-107
• /
• 2021

This paper discusses the containment control problem for multi-agent systems with input saturations. The goal of the containment control is to obtain swarming behavior by driving follower agents into the convex hull which is spanned by multiple leader agents. This paper considers multiple leader agents moving at the same constant speed. Then, to solve the containment problem for moving leaders, we propose a PI-based distributed control algorithm. We next analyze the convergence of follower agents to the desired positions. Specifically, we apply the integral-type Lyapunov function to take into account the saturation nonlinearity. Then, based on Lasalle's Invariance Principle, we show that the asymptotic convergence of error states to zero for any positive constant gains. Finally, numerical examples with the static and moving leaders are provided to validate the theoretical results.

• The KIPS Transactions:PartD
• /
• v.11D no.1
• /
• pp.105-114
• /
• 2004

In this paper, we propose a collaboration method of diverse agents each others in multi-agent model and describe a node migration algorithm of Mobile-Agent (MA) using by the metadata of Naming-Agent (NA). Collaboration work of multi-agent assures stability of agent system and provides reliability of information retrieval on the distributed environment. NA, an important part of multi-agent, identifies each agents and series the unique name of each agents, and each agent references the specified object using by its name. Also, NA integrates and manages naming service by agents classification such as Client-Push-Agent (CPA), Server-Push-Agent (SPA), and System-Monitoring-Agent (SMA) based on its characteristic. And, NA provides the location list of mobile nodes to specified MA. Therefore, when MA does move through the nodes, it is needed to improve the efficiency of node migration by specified priority according to hit_count, hit_ratio, node processing and network traffic time. Therefore, in this paper, for the integrated naming service, we design Naming Agent and show the structure of metadata which constructed with fields such as hit_count, hit_ratio, total_count of documents, and so on. And, this paper presents the flow of creation and updating of metadata and the method of node migration with hit_count through the collaboration of multi-agent.