Filters








1,503 Hits in 10.8 sec

Towards an Optimal Scoring Policy for Simulated Soccer Agents [chapter]

Jelle Kok, Remco de Boer, Nikos Vlassis, Frans C. A. Groen
2003 Lecture Notes in Computer Science  
This paper describes the implementation of a scoring policy that was used by the agents of the UvA Trilearn 2001 soccer simulation team during the RoboCup-2001 robotic soccer world championship.  ...  In a given situation this policy enables agents to determine the best shooting point in the goal, together with an associated probability of scoring when the ball is shot to this point.  ...  In the remainder of this paper we describe the implementation of a scoring policy that was used by the agents of the UvA Trilearn 2001 soccer simulation team during the RoboCup-2001 robotic soccer world  ... 
doi:10.1007/978-3-540-45135-8_24 fatcat:azliupof4zf7vkdpgwv465fkja

Reinforcement Learning for Soccer Multi-agents System

Fahimeh Farahnakian, Nasser Mozayani
2009 2009 International Conference on Computational Intelligence and Security  
Therefore we use reinforcement learning for optimizing policy.  ...  Experimental results have shown that policy achieved from reinforcement learning lead to more effective shoots toward the goal in simulated soccer agent.  ...  Behavior learning for complex tasks is also an important research area in RoboCup. In our previous describes the use of decision tree to kick and catch the ball for two simulated soccer agents [4] .  ... 
doi:10.1109/cis.2009.275 dblp:conf/cis/FarahnakianM09 fatcat:cmwwbaxufzdqdgxstjgazrntg4

Integrating learning with motor schema-based control for a Robot Soccer Team [chapter]

Tucker Balch
1998 Lecture Notes in Computer Science  
This paper describes a reinforcement learning-based strategy developed for Robocup simulator league competition.  ...  The entire team is jointly rewarded or penalized when they score or are scored against (global reinforcement). This approach provides for diversity in individual behavior.  ...  If the function is properly computed, an agent can act optimally simply by looking up the bestvalued action for any situation. The problem is to find the Q(s, a) that provides an optimal policy.  ... 
doi:10.1007/3-540-64473-3_86 fatcat:4houy65jird6fovs2kja6i44p4

CBR for State Value Function Approximation in Reinforcement Learning [chapter]

Thomas Gabel, Martin Riedmiller
2005 Lecture Notes in Computer Science  
The approach we take is evaluated in a case study in robotic soccer simulation.  ...  In Reinforcement Learning systems a learning agent is faced with the problem of assessing the desirability of the state it finds itself in.  ...  soccer simulation system.  ... 
doi:10.1007/11536406_18 fatcat:rwxynjvn4nc4hl7mgngfhdtxjq

Ant Intelligence in Robotic Soccer

R. Geetha Ramani, P. Viswanath, B. Arjun
2008 International Journal of Advanced Robotic Systems  
The simulation team evolved (PUTeam) was tested with teams of soccerbots in teambots (a simulation tool for Robotic Soccer) and the experimental results clearly shows the performance of the evolved team  ...  Robotic Soccer is a multi-agent test bed, which requires the designer to address most of the issues of multi-agent research.  ...  If the function is properly computed, an agent can act optimally simply by looking up the best valued action for any situation. The problem is to find the Q(s, a) s that provides an optimal policy.  ... 
doi:10.5772/5657 fatcat:r77wjfjs4rhqtmun4vei3jkxf4

A Framework for Studying Reinforcement Learning and Sim-to-Real in Robot Soccer [article]

Hansenclever F. Bassani, Renie A. Delgado, José Nilton de O. Lima Junior, Heitor R. Medeiros, Pedro H. M. Braga, Mateus G. Machado, Lucas H. C. Santos, Alain Tapp
2020 arXiv   pre-print
This article introduces an open framework, called VSSS-RL, for studying Reinforcement Learning (RL) and sim-to-real in robot soccer, focusing on the IEEE Very Small Size Soccer (VSSS) league.  ...  We propose a simulated environment in which continuous or discrete control policies can be trained to control the complete behavior of soccer agents and a sim-to-real method based on domain adaptation  ...  Cientifico e Tecnológico (CNPq), and Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES) for financial support.  ... 
arXiv:2008.12624v1 fatcat:27dc3ukzofhtzmr7mrgbp7icua

Multiagent Reinforcement Learning with Regret Matching for Robot Soccer

Qiang Liu, Jiachen Ma, Wei Xie
2013 Mathematical Problems in Engineering  
and results in significant performance in terms of scores, average reward and policy convergence.  ...  Simulation results on robot soccer validate that compared to original Nash-learning algorithm, the use of regret matching during the learning phase of Nash-learning has excellent ability of online learning  ...  An extended simulation shows that the average number of policy changes for Team A reached zero after 150 matches.  ... 
doi:10.1155/2013/926267 fatcat:h5uwpcaom5anhiadpihh5xhvji

rSoccer: A Framework for Studying Reinforcement Learning in Small and Very Small Size Robot Soccer [article]

Felipe B. Martins, Mateus G. Machado, Hansenclever F. Bassani, Pedro H. M. Braga, Edna S. Barros
2021 arXiv   pre-print
This article introduces an open-source simulator for the IEEE Very Small Size Soccer and the Small Size League optimized for reinforcement learning experiments.  ...  We also propose a framework for creating OpenAI Gym environments with a set of benchmarks tasks for evaluating single-agent and multi-agent robot soccer skills.  ...  Cientifico e Tecnológico (CNPq), and Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES) for financial support.  ... 
arXiv:2106.12895v1 fatcat:o5oq5jmz7ngilkmg26zfbjaom4

Emergent Coordination Through Competition [article]

Siqi Liu, Guy Lever, Josh Merel, Saran Tunyasuvunakool, Nicolas Heess, Thore Graepel
2019 arXiv   pre-print
We study the emergence of cooperative behaviors in reinforcement learning agents by introducing a challenging competitive multi-agent soccer environment with continuous simulated physics.  ...  We further apply an evaluation scheme, grounded by game theoretic principals, that can assess agent performance in the absence of pre-defined evaluation tasks or human baselines.  ...  This trend is subsequently reversed towards the end of training, where the agents evolved to pay more attention to conceding goals: i.e. agents first learn to optimize scoring and then incorporate defending  ... 
arXiv:1902.07151v2 fatcat:dbhxgyk5lzb4zhhoflvjt642sy

Towards collaborative and adversarial learning: a case study in robotic soccer

PETER STONE, MANUELA VELOSO
1998 International Journal of Human-Computer Studies  
Soccer is a rich domain for the study of multiagent learning issues.  ...  We are using a robotic soccer system to study these different types of multiagent learning: low-level skills, collaborative, and adversarial. Here we describe in detail our experimental framework.  ...  The novice learns to drive on a simulated race track from an expert agent whose behavior is fixed.  ... 
doi:10.1006/ijhc.1997.0162 fatcat:vr7iwy4ujjebnhzvjajzuzyazy

Evaluating the impact of curriculum learning on the training process for an intelligent agent in a video game

Jorge E Camargo, Rigoberto Sáenz
2021 Inteligencia Artificial  
Then all were executed on a selected game simulation platform, using two reinforcement learning algorithms, and using the mean cumulative reward as a performance measure.  ...  We want to measure the impact of the curriculum learning technique on a reinforcement training setup, several experiments were designed with different training curriculums adapted for the video game chosen  ...  found appropriate to train our agents: Proximal Policy Optimization (PPO) [30] .  ... 
doi:10.4114/intartif.vol24iss68pp1-20 fatcat:sjfcptixn5cedkytvfdfwavl3i

A Case Study on Improving Defense Behavior in Soccer Simulation 2D: The NeuroHassle Approach [chapter]

Thomas Gabel, Martin Riedmiller, Florian Trost
2009 Lecture Notes in Computer Science  
We employ a reinforcement learning methodology that enables our players to autonomously acquire such an aggressive duel behavior, and we have embedded it into our soccer simulation team's defensive strategy  ...  While a lot of papers on RoboCup's robotic 2D soccer simulation have focused on the players' offensive behavior, there are only a few papers that specifically address a team's defense strategy.  ...  Therefore, while the reward function is unknown to the agent, in soccer simulation the transition function p (model of the environment) is given since the way the Soccer Server simulates a soccer match  ... 
doi:10.1007/978-3-642-02921-9_6 fatcat:yzqi2jqlffbovk2jszniu7tudi

Cooperative reinforcement learning based on zero-sum games

Kao-Shing Hwang, Jeng-Yih Chiou, Tse-Yu Chen
2008 2008 SICE Annual Conference  
Although centralized optimization provides opportunity for efficient optimization, the coordination and the transfer of information among agents are costly and often infeasible.  ...  Hence it is important to develop decentralized optimization schemes which permit the individual agents take control of the actions that contribute towards the optimization of a global performance criteria  ...  Cooperative path planning, formation control of multi robotic agents, communication and distance measurement between agents are shown.  ... 
doi:10.1109/sice.2008.4655172 fatcat:vgnliviuvnacdg3rtdrrii23eu

Cooperative Reinforcement Learning Based on Zero-Sum Games [chapter]

Kao-Shing Hwang, Wei-Cheng Jiang, Hung-Hsiu Yu, Shin-Yi Li
2011 Mobile Robots - Control Architectures, Bio-Interfacing, Navigation, Multi Robot Motion Planning and Operator Training  
Although centralized optimization provides opportunity for efficient optimization, the coordination and the transfer of information among agents are costly and often infeasible.  ...  Hence it is important to develop decentralized optimization schemes which permit the individual agents take control of the actions that contribute towards the optimization of a global performance criteria  ...  Cooperative path planning, formation control of multi robotic agents, communication and distance measurement between agents are shown.  ... 
doi:10.5772/26620 fatcat:u53pcolpmrdhnmxayihlgiqrxm

On Experiences in a Complex and Competitive Gaming Domain: Reinforcement Learning Meets RoboCup

Martin Riedmiller, Thomas Gabel
2007 2007 IEEE Symposium on Computational Intelligence and Games  
RoboCup soccer simulation features the challenges of a fully distributed multi-agent domain with continuous state and action spaces, partial observability, as well as noisy perception and action execution  ...  While the application of machine learning techniques in this domain represents a promising idea in itself, the competitive character of RoboCup also evokes the desire to head for the development of learning  ...  In its pure form, this corresponds to a policy iteration step which necessarily leads to an improvement of the policy until the optimal policy is found.  ... 
doi:10.1109/cig.2007.368074 dblp:conf/cig/RiedmillerG07 fatcat:3xp2mrgi7fc5rcsnigclxovksa
« Previous Showing results 1 — 15 out of 1,503 results