• 제어로봇시스템학회:학술대회논문집
• /
• 2001.10a
• /
• pp.149.5-149
• /
• 2001

The robot soccer simulation game is a dynamic multi-agent environment. In this paper we suggest a new reinforcement learning approach to each agent´s dynamic positioning in such dynamic environment. Reinforcement learning is the machine learning in which an agent learns from indirect, delayed reward an optimal policy to choose sequences of actions that produce the greatest cumulative reward. Therefore the reinforcement learning is different from supervised learning in the sense that there is no presentation of input-output pairs as training examples. Furthermore, model-free reinforcement learning algorithms like Q-learning do not require defining or learning any models of the surrounding environment. Nevertheless ...

• Journal of Korea Game Society
• /
• v.15 no.5
• /
• pp.131-142
• /
• 2015

A self-driving car is an autonomous vehicle capable of fulfilling the main transportation capabilities of a traditional car. It must be capable of sensing its environment and navigating without human input. In this paper, we design the agent that can simulate these self-driving cars and develop a prototype for it. To do this, we analyze the requirements for the self-driving car, and then the agent is designed to be suitable for traditional multi-agent system. The key point of the design is that agents move along the steering forces only. The prototype of the designed agent was implemented by using Unity 3D. From simulation results using the prototype, movements of the agents were very realistic. However, in the case of increasing the number of the agent the performance was seriously degraded, and so the alternatives of the problem were suggested.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.31 no.5B
• /
• pp.405-412
• /
• 2006

In this paper, we present an efficient method for simulating massively virtual clients in an online game environment. Massively multi-player online games and other multi-user based networked applications are becoming more attractive to the gamer players. Such kind of technology has long been researched in the area called Networked Virtual Environments. In the game development process, a set of beta tests is used to ensure the stability of online game servers. A set of testing processes consumesa lot of development resources such as cost, time, and etc. The purpose of VENUS system is to provide an automated beta test environment to the game developers to efficiently test the online games to reduce development resources.

• International Journal of Computer Science & Network Security
• /
• v.23 no.8
• /
• pp.159-167
• /
• 2023

Objective: • To detect black hole and warm hole attacks in wireless sensor networks. • To give a solution for energy depletion and security breach in wireless sensor networks. • To address the security problem using strategic decision support system. Methods: The proposed stackelberg game is used to make the spirited relations between multi leaders and multi followers. In this game, all cluster heads are acts as leaders, whereas agent nodes are acts as followers. The game is initially modeled as Quadratic Programming and also use backtracking search optimization algorithm for getting threshold value to determine the optimal strategies of both defender and attacker. Findings: To find optimal payoffs of multi leaders and multi followers are based on their utility functions. The attacks are easily detected based on some defined rules and optimum results of the game. Finally, the simulations are executed in matlab and the impacts of detection of black hole and warm hole attacks are also presented in this paper. Novelty: The novelty of this study is to considering the stackelberg game with backtracking search optimization algorithm (BSOA). BSOA is based on iterative process which tries to minimize the objective function. Thus we obtain the better optimization results than the earlier approaches.

• Journal of Korea Game Society
• /
• v.20 no.1
• /
• pp.57-66
• /
• 2020

The goal is to implement AI using reinforcement learning, which replaces the less favored Supporter in MOBA games. ML_Agent implements game rules, environment, observation information, rewards, and punishment. The experiment was divided into P and C group. Experiments were conducted to compare the cumulative compensation values and the number of deaths to draw conclusions. In group C, the mean cumulative compensation value was 3.3 higher than that in group P, and the total mean number of deaths was 3.15 lower. performed cooperative play to minimize death and maximize rewards was confirmed.

• Proceedings of the Korean Society of Precision Engineering Conference
• /
• 1996.11a
• /
• pp.434-438
• /
• 1996

In this paper we present the multi-agent robot system and the vision system developed for participating in micro robot soccer tournament. The multi-agent robot system consists of micro robot, a vision system, a host computer and a communication module. Micro robot are equipped with two mini DC motors witf encoders and gearboxes, a R/F receiver, a CPU and infrared sensors for obstacle detection. A vision system is used to recognize the position of the ball and opponent robots, position and orientation of our robots. The vision system is composed of a color CCD camera and a vision processing unit(AISI vision computer). The vision algorithm is based on morphological method. And it takes about 90 msec to detect ball and 3-our robots and 3-opponent robots with reasonable accuracy

• Journal of Korea Multimedia Society
• /
• v.10 no.12
• /
• pp.1714-1725
• /
• 2007

This paper describes a top-down behavior planning framework in a simulation game from personality to real life action selection. The combined behavior creating system is formed by five levels of specification, which are personality definition, motivation extraction, emotion generation, decision making and action execution. Along with the data flowing process in our designed framework, NPC selects actions autonomously to adapt to the dynamic environment information resulting from active agents and human players. Furthermore, we illuminate applying Gaussian probabilistic distribution to realize character's behavior changeability like human performance. To elucidate the mechanism of the framework, we situated it in a restaurant simulation game.

• Journal of Internet Computing and Services
• /
• v.7 no.2
• /
• pp.137-149
• /
• 2006

In this paper, we suggest a message-based communication protocol and a dynamic role assignment mechanism for efficient cooperation and teamwork of agents in computer game environments. This role assignment mechanism is a new one different from both existing static and dynamic mechanisms, in which all decisions related with role assignment are made at once in design phase or execution phase. According to our mechanism, all possible role sets are determined in design phase. Detail decisions regarding which agent takes what role, however, are made in execution phase. This mechanism for role assignment can minimize the negotiation effort in execution phase. Therefore, this mechanism is quite effective especially in real-time multiagent environments. Through experiments, we show the superiority of the new dynamic role assignment mechanism.

• Journal of the Korea Society of Computer and Information
• /
• v.14 no.12
• /
• pp.33-40
• /
• 2009

We study Interpreted Systems, ATL, and ATEL to capture the notion of time, knowledge, and strategy which are important in the analysis of multi-agent systems and propose strategic constraints based on subgame perfect Nash equilibrium of game theory as one of the solutions for the issues of ATEL which an agent can access the current state of the whole system when making up his strategy even when he should be uncertain about the state, and no explicit representation of actions in ATEL models makes some natural situations harder to model. Also, we present strategic constraints-based Interpreted Systems for model checking of multi-agent systems.

• 제어로봇시스템학회:학술대회논문집
• /
• 1997.10a
• /
• pp.612-616
• /
• 1997

In this paper, two kinds of controller are proposed for a soccer robot system.. One for Supervisor and defense mode, and the other for attack mode. Robot soccer game has very dynamic characteristics. Furthermore, there exist competitions between agents. The soccer-playing robot should take an appropriate action according to its surroundings. Initially, an attack mode controller using a vector field concept is designed, then a supervisor and a defense mode controller are designed with a Petri-net. The efficiency and applicability of the proposed controllers are demonstrated through a real robot soccer game(MiroSot 97).