• 한국정보과학회논문지:컴퓨팅의 실제 및 레터
• /
• 제14권5호
• /
• pp.502-506
• /
• 2008

최근에 제안된 강화 학습 기법인 "potential-based" reinforcement learning(RL) 기법은 다수 학습들과 expert advice들을 감독 지식으로 강화 학습 알고리즘에 융합하는 것을 가능하게 했고 그 효용성은 최적 정책으로의 이론적 수렴성 보장으로 증명되었다. 본 논문에서는 potential-based RL 기법을 셀룰러 네트워크에서의 채널 할당 문제에 적용한다. Potential-based RL 기반의 동적 채널 할당 기법이 기존의 fixed channel assignment, Maxavail, Q-learning-based dynamic channel assignment 채널 할당 기법들보다 효율적으로 채널을 할당한다. 또한, potential-based RL 기법이 기존의 강화 학습 알고리즘인 Q-learning, SARSA(0)에 비하여 최적 정책에 더 빠르게 수렴함을 실험적으로 보인다.

• Journal of Information Processing Systems
• /
• 제16권5호
• /
• pp.1048-1063
• /
• 2020

In this study, we propose a reinforcement learning agent to control the data transmission rates of nodes in carrier sensing multiple access with collision avoidance (CSMA/CA)-based wireless networks. We design a reinforcement learning (RL) agent, based on Q-learning. The agent learns the environment using the timeout events of packets, which are locally available in data sending nodes. The agent selects actions to control the data transmission rates of nodes that adjust the modulation and coding scheme (MCS) levels of the data packets to utilize the available bandwidth in dynamically changing channel conditions effectively. We use the ns3-gym framework to simulate RL and investigate the effects of the parameters of Q-learning on the performance of the RL agent. The simulation results indicate that the proposed RL agent adequately adjusts the MCS levels according to the changes in the network, and achieves a high throughput comparable to those of the existing data transmission rate adaptation schemes such as Minstrel.

• 한국정보과학회논문지:컴퓨팅의 실제 및 레터
• /
• 제15권12호
• /
• pp.978-982
• /
• 2009

본 논문에서는 다수의 강화 학습 에이전트들의 학습 결과 및 Expert의 지식을 하나의 학습 알고리즘으로 융합하는 강화학습인 "potential-based" reinforcement learning(RL)기법에 불확실한 환경에서의 의사결정 알고리즘인 Case-based Decision Theory(CBDT)를 적용한 "RLs-CBDT"를 제안한다. 그리고 테트리스 실험을 통하여 기존의 RL 알고리즘에 비해 RLs-CBDT가 최적의 정책에 더 마르게 수렴하는 것을 보인다.

• International Journal of Internet, Broadcasting and Communication
• /
• 제16권1호
• /
• pp.63-71
• /
• 2024

Deep reinforcement learning (RL) significantly influences autonomous vehicle development by optimizing decision-making and adaptation to complex driving environments through simulation-based training. In deep RL, an activation function is used, and various activation functions have been proposed, but their performance varies greatly depending on the application environment. Therefore, finding the optimal activation function according to the environment is important for effective learning. In this paper, we analyzed nine commonly used activation functions for RL to compare and evaluate which activation function is most effective when using deep RL for autonomous vehicles to learn highway merging. To do this, we built a performance evaluation environment and compared the average reward of each activation function. The results showed that the highest reward was achieved using Mish, and the lowest using SELU. The difference in reward between the two activation functions was 10.3%.

• Nuclear Engineering and Technology
• /
• 제56권6호
• /
• pp.1959-1974
• /
• 2024

Nuclear energy plays a crucial role in energy supply in the 21st century, and more and more Nuclear Power Plants (NPPs) will be in operation to contribute to the development of human society. However, as a typical complex system engineering, the operation and development of NPPs require efficient and stable control methods to ensure the safety and efficiency of nuclear power generation. Reinforcement learning (RL) aims at learning optimal control policies via maximizing discounted long-term rewards. The reward-oriented learning paradigm has witnessed remarkable success in many complex systems, such as wind power systems, electric power systems, coal fire power plants, robotics, etc. In this work, we try to present a systematic review of the applications of RL on these complex systems, from which we believe NPPs can borrow experience and insights. We then conduct a block-by-block investigation on the application scenarios of specific tasks in NPPs and carried out algorithmic research for different situations such as power startup, collaborative control, and emergency handling. Moreover, we discuss the possibilities of further application of RL methods on NPPs and detail the challenges when applying RL methods on NPPs. We hope this work can boost the realization of intelligent NPPs, and contribute to more and more research on how to better integrate RL algorithms into NPPs.

• International Journal of Computer Science & Network Security
• /
• 제21권9호
• /
• pp.125-131
• /
• 2021

Detecting fraudulent insurance claims is difficult due to small and unbalanced data. Some research has been carried out to better cope with various types of fraudulent claims. Nowadays, technology for detecting fraudulent insurance claims has been increasingly utilized in insurance and technology fields, thanks to the use of artificial intelligence (AI) methods in addition to traditional statistical detection and rule-based methods. This study obtained meaningful results for a fraudulent insurance claim detection model based on machine learning (ML) and deep learning (DL) technologies, using fraudulent insurance claim data from previous research. In our search for a method to enhance the detection of fraudulent insurance claims, we investigated the reinforcement learning (RL) method. We examined how we could apply the RL method to the detection of fraudulent insurance claims. There are limited previous cases of applying the RL method. Thus, we first had to define the RL essential elements based on previous research on detecting anomalies. We applied the deep Q-network (DQN) and double deep Q-network (DDQN) in the learning fraudulent insurance claim detection model. By doing so, we confirmed that our model demonstrated better performance than previous machine learning models.

• International Journal of Computer Science & Network Security
• /
• 제24권4호
• /
• pp.87-106
• /
• 2024

Vehicle-to-Home (V2H) and Home Centralized Photovoltaic (HCPV) systems can address various energy storage issues and enhance demand response programs. Renewable energy, such as solar energy and wind turbines, address the energy gap. However, no energy management system is currently available to regulate the uncertainty of renewable energy sources, electric vehicles, and appliance consumption within a smart microgrid. Therefore, this study investigated the impact of solar photovoltaic (PV) panels, electric vehicles, and Micro-Grid (MG) storage on maximum solar radiation hours. Several Deep Learning (DL) algorithms were applied to account for the uncertainty. Moreover, a Reinforcement Learning HCPV (RL-HCPV) algorithm was created for efficient real-time energy scheduling decisions. The proposed algorithm managed the energy demand between PV solar energy generation and vehicle energy storage. RL-HCPV was modeled according to several constraints to meet household electricity demands in sunny and cloudy weather. Simulations demonstrated how the proposed RL-HCPV system could efficiently handle the demand response and how V2H can help to smooth the appliance load profile and reduce power consumption costs with sustainable power generation. The results demonstrated the advantages of utilizing RL and V2H as potential storage technology for smart buildings.

• 인터넷정보학회논문지
• /
• 제22권4호
• /
• pp.13-28
• /
• 2021

강화학습은 한 환경에서 에이전트가 정책에 따라 액션을 취하고 보상 함수를 통해 액션을 평가 및 정책 최적화 과정을 반복하는 Closed-Loop 구조로 이루어진 알고리즘이다. 이러한 강화학습의 주요 장점은 액션의 품질을 평가하고 정책을 지속적으로 최적화 하는 것이다. 따라서, 강화학습은 지능형 시스템, 자율제어 시스템 개발에 효과적으로 활용될 수 있다. 기존의 강화학습은, 단일 정책, 단일 보상함수 및 비교적 단순한 정책 업데이트 기법을 제한적인 문제에 대해 제시하고 적용하였다. 본 논문에서는 구성요소의 복수성을 지원하는 확장된 강화학습 모델을 제안한다. 제안되는 확정 강화학습의 주요 구성 요소들을 정의하고, 그들의 컴퓨팅 모델을 포함하는 정형 모델을 제시한다. 또한, 이 정형모델을 기반으로 시스템 개발을 위한 설계 기법을 제시한다. 제안한 모델을 기반으로 자율 최적화 자동차 내비게이터 시스템에 적용 및 실험을 진행한다. 제시된 정형 모델과 설계 기법을 적용한 사례연구로, 복수의 자동차들이 최적 목적지에 단 시간에 도착할 수 있는 진화된 내비게이터 시스템 설계 및 구현을 진행한다.

• 정보처리학회논문지:소프트웨어 및 데이터공학
• /
• 제10권7호
• /
• pp.263-270
• /
• 2021

강화학습은 학습을 통해 최적의 행동정책을 탐색하는 기법으로써, 재난 상황에서 효과적인 인명 구조 및 재난 대응 문제 해결을 위해 많이 활용되고 있다. 그러나, 기존 재난 대응을 위한 강화학습 기법은 상대적으로 단순한 그리드, 그래프와 같은 환경 혹은 자체 개발한 강화학습 환경을 통해 평가를 수행함에 따라 그 실용성이 충분히 검증되지 않았다. 본 논문에서는 강화학습 기법을 실세계 환경에서 사용하기 위해 기존 개발된 재난 시뮬레이션 환경의 복잡한 프로퍼티를 활용하는 강화학습 환경 구성과 활용 결과를 제시하고자 한다. 본 제안 강화학습 환경의 구성을 위하여 재난 시뮬레이션과 강화학습 에이전트 간 강화학습 커뮤니케이션 채널 및 인터페이스를 구축하였으며, 시뮬레이션 환경이 제공하는 고차원의 프로퍼티 정보의 활용을 위해 비-이미지 피쳐 벡터(non-image feature vector)에 이미지 변환방식을 적용하였다. 실험을 통해 본 제안 방식이 건물 화재 피해도를 기준으로 한 평가에서 기존 방식 대비 가장 낮은 건물 화재 피해를 기록한 것을 확인하였다.

• 한국정보기술학회 영문논문지
• /
• 제9권1호
• /
• pp.127-134
• /
• 2019

In terms of deep Reinforcement Learning (RL), exploration can be worked stochastically in the action of a state space. On the other hands, exploitation can be done the proportion of well generalization behaviors. The balance of exploration and exploitation is extremely important for better results. The randomly selected action with ε-greedy for exploration has been regarded as a de facto method. There is an alternative method to add noise parameters into a neural network for richer exploration. However, it is not easy to predict or detect over-fitting with the stochastically exploration in the perturbed neural network. Moreover, the well-trained agents in RL do not necessarily prevent or detect over-fitting in the neural network. Therefore, we suggest a novel design of a deep RL by the balance of the exploration with drop-out to reduce over-fitting in the perturbed neural networks.