On Learning Soccer
نویسندگان
چکیده
We use simulated soccer to study multiagent learning. Each team's players (agents) share action set and policy but may behave differently due to position-dependent inputs. All agents making up a team are rewarded or punished collectively in case of goals. We conduct simulations with varying team sizes, and compare two learning algorithms: TD-Q learning with linear neural networks (TD-Q) and Probabilistic Incremental Program Evolution (PIPE). TD-Q is based on evaluation functions (EFs) mapping input/action pairs to expected reward, while PIPE searches policy space directly. PIPE uses an adaptive probability distribution to synthesize programs that calculate action probabilities from current inputs. Our results show that TD-Q has diiculties to learn appropriate shared EFs. PIPE, however, does not depend on EFs and nds good policies faster and more reliably.
منابع مشابه
An Unsupervised Learning Method for an Attacker Agent in Robot Soccer Competitions Based on the Kohonen Neural Network
RoboCup competition as a great test-bed, has turned to a worldwide popular domains in recent years. The main object of such competitions is to deal with complex behavior of systems whichconsist of multiple autonomous agents. The rich experience of human soccer player can be used as a valuable reference for a robot soccer player. However, because of the differences between real and simulated soc...
متن کاملSoccer Goalkeeper Task Modeling and Analysis by Petri Nets
In a robotic soccer team, goalkeeper is an important challenging role, which has different characteristics from the other teammates. This paper proposes a new learning-based behavior model for a soccer goalkeeper robot by using Petri nets. The model focuses on modeling and analyzing, both qualitatively and quantitatively, for the goalkeeper role so that we have a model-based knowledge of the ta...
متن کاملLearning in Robot Soccer
This paper presents five case studies of successful applications of reinforcement learning methods to learn specific subtasks in robot soccer. The used learning algorithms are introduced, followed by descriptions of each case study on its own, by introducing the task, the machine learning techniques and finally the results. The case studies range from motor control over simulated soccer to actu...
متن کاملLearning Strategies for Coordination of Multi Robot Systems: a Robot Soccer Application
This paper presents a hybrid method for learning a dynamic strategy for a robot soccer team. In this method, an imitation learning scheme based on observed robot soccer games is used as a seed for an experience-guided learning scheme based on reinforcement learning. A lack in the application of classic reinforcement learning to the robot soccer problem is the high number of states to be analyze...
متن کاملEvent Detection Based Approach for Soccer Video Summarization Using Machine learning
Many soccer fans prefer to watch a summary of football games as watching a whole soccer match needs a lot of time. Traditionally, soccer videos were analyzed manually, however this costs valuable time. Therefore, it is necessary to have a tool for doing the video analysis and summarization job automatically. Automatic soccer video summarization is about extracting important events from soccer m...
متن کاملThe key to success in elite athletes? Explicit and implicit motor learning in youth elite and non-elite soccer players.
In sports, fast and accurate execution of movements is required. It has been shown that implicitly learned movements might be less vulnerable than explicitly learned movements to stressful and fast changing circumstances that exist at the elite sports level. The present study provides insight in explicit and implicit motor learning in youth soccer players with different expertise levels. Twenty...
متن کامل