• 한국시스템다이내믹스연구
• /
• 제5권2호
• /
• pp.175-197
• /
• 2004

The research studies the classic beer game simulation model from a new perspective. It does so by providing each agent with two ordering policies, and creating a set of rules that allow an agent to change its policy. Such a change is triggered based on an agent's confidence in their own performance, and on the relative confidence of their nearest neighbour. The overall effect is that policy diffusion can occur, where, under certain circumstances, an agent will mimic the behaviour of its neighbour, if it believes that its neighbour is performing better. The motivation behind this research is to provide an experimental base upon which the decision making strategies of business agent can be studied.

• 한국게임학회 논문지
• /
• 제19권6호
• /
• pp.61-70
• /
• 2019

본 논문에서는 경영 시뮬레이션 게임 분야에서 강화학습을 적용하여 게임 에이전트들이 자율적으로 주어진 목표를 달성하는지를 확인하고자 한다. 본 시스템에서는 Unity Machine Learning (ML) Agent 환경에서 PPO (Proximal Policy Optimization) 알고리즘을 적용하여 게임 에이전트가 목표를 달성하기 위해 자동으로 플레이 방법을 찾도록 설계하였다. 그 유용성을 확인하기 위하여 5가지의 게임 시나리오 시뮬레이션 실험을 수행하였다. 그 결과 게임 에이전트가 다양한 게임 내 환경 변수의 변화에도 학습을 통하여 목표를 달성한다는 것을 확인하였다.

• 한국정밀공학회:학술대회논문집
• /
• 한국정밀공학회 1994년도 추계학술대회 논문집
• /
• pp.670-675
• /
• 1994

Yhis paper describes a fault diagnosis simulation of the Real-Time Multiple Fault Dignosis System (RTMFDS) for forcasting faults in a system and deciding current machine state from signal information. Comparing with other diagnosis system for single fault,the system developed deals with multiple fault diagnosis,comprising two main parts. One is a remotesignal generating and transimission terminal and the other is a host system for fault diagnosis. Signal generator generate the random fault signal and the image information, and send this information to host. Host consists of various modules and agents such as Signal Processing Module(SPM) for sinal preprocessing, Performence Monotoring Module(PMM) for subsystem performance monitoring, Trigger Module(TM) for multi-triggering subsystem fault diagnosis, Subsystem Fault Diagnosis Agent(SFDA) for receiving trigger signal, formulating subsystem fault D\ulcornerB and initiating diagnosis, Fault Diagnosis Module(FDM) for simulating component fault with Hierarchical Artificial Neural Network (HANN), numerical models and Hofield network,Result Agent(RA) for receiving simulation result and sending to Treatment solver and Graphic Agent(GA). Each agent represents a separate process in UNIX operating system, information exchange and cooperation between agents was doen by IPC(Inter Process Communication : message queue, semaphore, signal, pipe). Numerical models are used to deseribe structure, function and behavior of total system, subsystems and their components. Hierarchical data structure for diagnosing the fault system is implemented by HANN. Signal generation and transmittion was performed on PC. As a host, SUN workstation with X-Windows(Motif)is used for graphic representation.

• International Journal of Computer Science & Network Security
• /
• 제21권7호
• /
• pp.297-304
• /
• 2021

In a duopoly situation, one firm can gain competitive advantage by attracting the price-sensitive customers from its competitor leading to higher profits through higher sale revenue for the competing company. A simulation study in which there are two electric car manufacturers with agent based modelling was conducted in order to verify this contention. The first step consisted of defining the baseline. Simulations of 1000 times and agent-based modelling were conducted with the assumption that company 1 reduced its price to the maximum of 20% thereby contributing to the switch-over of a maximum of 40% of the price sensitive customers of company 2. The results of 1000 simulations and agent-based modelling highlighted that price reduction by company 1 resulted in a significant increase in the number of customers, presumably due to switch-over from company 2 and there was a corresponding increase in revenues from both of the sales avenues. Thus, Company 1 achieved competitive advantage by marketing its cars to the customers of Company 2 using price reduction strategy to attract them. This study has ramifications for companies that aim to sway the price sensitive customers from a competitor.

• 한국시뮬레이션학회논문지
• /
• 제22권4호
• /
• pp.49-55
• /
• 2013

최근 M&S 및 DM&S (Defense M&S) 환경의 복잡성이 높아짐에 따라 인간 등의 자율세력을 표현하기 위한 기술로 CGF/SAF (Computer Generated Force/Semi-Automated Force)가 주목 받고 있다. 그러나 기존의 OneSAF 과 같은 선진국의 CGF/SAF 기반 DM&S 환경은 DEVS 프레임워크 중심의 국내 DM&S 시스템과 유기적 연동이 어렵다. 따라서 본 논문에서는 DEVS M&S환경을 지원하는 에이전트 기반의 SAF 설계 방법론 및 모델링 도구를 제안하였다. 제안한 SAF 모델링 도구는 에이전트 모델링 GUI와 SAF 모델링 GUI로 나누어진다. 에이전트 모델링 환경은 간단한 UI를 활용한 규칙 입력만으로 에이전트를 생성할 수 있다. 또한 에이전트 단위테스트 환경을 추가하여 UI 기반의 에이전트 모델의 검증이 가능하다. SAF 모델을 생성하기 위한 환경으로 개개의 에이전트 모델들의 구조적 결합을 정의할 수 있도록 하였다. 본 연구의 모델링 도구 및 방법론을 통해 DM&S 연구자 및 개발자들이 DEVS 기반의 DM&S 및 SAF 시스템을 구축하는데 유용하게 활용될 수 있을 것으로 기대된다.

• 한국시뮬레이션학회:학술대회논문집
• /
• 한국시뮬레이션학회 2001년도 The Seoul International Simulation Conference
• /
• pp.321-324
• /
• 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 Beaming 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 loaming is different from supervised teaming in the sense that there is no presentation of input-output pairs as training examples. Furthermore, model-free reinforcement loaming algorithms like Q-learning do not require defining or loaming any models of the surrounding environment. Nevertheless it can learn the optimal policy if the agent can visit every state-action pair infinitely. However, the biggest problem of monolithic reinforcement learning is that its straightforward applications do not successfully scale up to more complex environments due to the intractable large space of states. In order to address this problem, we suggest Adaptive Mediation-based Modular Q-Learning(AMMQL) as an improvement of the existing Modular Q-Learning(MQL). While simple modular Q-learning combines the results from each learning module in a fixed way, AMMQL combines them in a more flexible way by assigning different weight to each module according to its contribution to rewards. Therefore in addition to resolving the problem of large state space effectively, AMMQL can show higher adaptability to environmental changes than pure MQL. This paper introduces the concept of AMMQL and presents details of its application into dynamic positioning of robot soccer agents.

• 한국경영과학회:학술대회논문집
• /
• 한국경영과학회/대한산업공학회 2003년도 춘계공동학술대회
• /
• pp.808-813
• /
• 2003

Maintaining agility and responsiveness m designing and manufacturing activities are the key issues for manufacturing companies to cope with global competition. Distributed design and control systems are regarded as an efficient solution for agility and responsiveness. However, distributed nature of a manufacturing system complicates production activities such as design, simulation, scheduling, and execution control. Especially, existing simulation systems have limited external integration capabilities, which make it difficult to implement complex control mechanisms for the distributed manufacturing systems. Moreover, integration and coupling of heterogeneous components and models are commonly required for the simulation of complex distributed systems. In this paper, a collaborative and adaptive simulation architecture is proposed as an open framework for simulation and analysis of the distributed manufacturing enterprises. By incorporating agents with their distributed characteristics of autonomy, intelligence, and goal-driven behavior, the proposed agent-based simulation architecture can be easily adapted to support the agile and distributed manufacturing systems. The architecture supports the coordination and cooperation relations, and provides a communication middleware among the participants in simulation.

• 지능정보연구
• /
• 제6권1호
• /
• pp.113-128
• /
• 2000

This paper presents a first year report of an ongoing multi-year project to develop a systolic parallel simulation system for dynamic traffic assignment. The fundamental approach to the simulation is systolic parallel processing based on autonomous agent modeling. Agents continuously act on their own initiatives and access to database to get the status of the simulation world. Various agents are defined in order to populate the simulation world. In particular existing modls and algorithm were incorporated in designing the behavior of relevant agents such as car-following model headway distribution Frank-Wolf algorithm and so on. Simulation is based on predetermined routes between centroids that are computed off-line by a conventional optimal path-finding algorithm. Iterating the cycles of optimization-then-simulation the proposed system will provide a realistic and valuable traffic assignment. Gangnum-Gu district in Seoul is selected for the target are for the modeling. It is expected that realtime traffic assignment services can be provided on the internet within 3 years.

• 제어로봇시스템학회논문지
• /
• 제5권7호
• /
• pp.848-854
• /
• 1999

In this paper, we designed and implemented a precise pulse motor controller chip that generates the pulse needed to control step motor, DC servo and AC servo motors. This chip generates maximum pulse output rate of 5Mpps and has the quasi-S driving capability and speed and moving distance override capability during driving. We designed this chip with VHDL and executed a logic simulation and synthesis using Synopsys tool. The pre-layout simulation and post-layout simulation was executed by Compass tool. This chip was produced with 100 pins, PQFP package by 0.8${\mu}{\textrm}{m}$ gate array process and implemented by completely digital logic. We developed the test hardware board of performance and the CAMC(Computer Aided Motor Controller) Agent softwate to test the performance of the pulse motor controller chip produced. CAMC Agent enables user to set parameters needed to control motor with easy GUI(Graphic User Interface) environment and to display the output response of motor graphically.

• 한국시뮬레이션학회:학술대회논문집
• /
• 한국시뮬레이션학회 2003년도 춘계학술대회논문집
• /
• pp.91-100
• /
• 2003

In this paper we investigate the role that warfare played In the formation of the network of alliances between sites that are associated with the formation of the state in the Valley of Oaxaca, Mexico. A model of state formation proposed by Marcos and Flannery (1996) is used as the basis for an agent-based simulation model. Agents reside in sites and their actions are constrained by knowledge extracted from the Oaxaca Surface Archaeological Survey (Kowalewski 1989). The simulation is run with two different sets of constraint rules for the agents. The first set is based upon the raw data collected in the surface survey. This represents a total of 79 sites and constitutes a minimal level of warfare (raiding) in the Valley. The other site represents the generalization of these constraints to sites with similar locational characteristics. This set corresponds to 987 sites and represents a much more active role for warfare in the Valley. The rules were produced by a data mining technique, Decision Trees, guided by Genetic Algorithms. Simulations were run using the two different rule sets and compared with each other and the archaeological data for the Valley. The results strongly suggest that warfare was a necessary process in the aggregations of resources needed to support the emergence of the state in the Valley.