Learning from Recorded Games : A Scoring Policy for Simulated Soccer Agents

This paper outlines the implementation of a new scoring policy for the agents of the Simulated Robot Soccer team from the University of Koblenz, called RoboLog. The applied technique is capable of acting in real time in the dynamic environment of the RoboCup Simulation League and uses data obtained from prerecorded soccer games for supervised neural network learning. The benchmark used for testing this approach is the Optimal Scoring Problem stated as finding the point in the goal where the probability of scoring is the highest when the ball is shot to this point in a given situation. Goalshot situations from numerous logfiles are extracted and employed for the training of two independent multi layered perceptrons. Beside the usage as training patterns the gained data is evaluated statistically and provides interesting general insights into goalshots carried out lately in Simulated Robot Soccer. The results obtained after extensive testing of the new policy are presented. Furthermore, general issues of learning from observed logfile data and starting points for future work are discussed.

[1]  Jürgen Ebert,et al.  GraX-an interchange format for reengineering tools , 1999, Sixth Working Conference on Reverse Engineering (Cat. No.PR00303).

[2]  Stephan Philippi,et al.  Modelling a concurrent ray-tracing algorithm using object-oriented Petri-Nets , 2001 .

[3]  Peter Baumgartner,et al.  The Taming of the (X)OR , 2000, Computational Logic.

[4]  Jens Woch,et al.  Implementation of a Schema-TAG-Parser , 1999 .

[5]  Oliver Obst,et al.  Spatial Agents Implemented in a Logical Expressible Language , 1999, RoboCup.

[6]  Andy Schürr,et al.  GXL: toward a standard exchange format , 2000, Proceedings Seventh Working Conference on Reverse Engineering.

[7]  Guillermo R. Simari,et al.  lntroducing generalized specificity in logic programming , 2000 .

[8]  Milind Tambe,et al.  Towards Automated Team Analysis : A Machine Learning Approach , 1999 .

[9]  Oliver Obst.  Specifying Rational Agents with Statecharts and Utility Functions , 2001, RoboCup.

[10]  Gerhard Weiss,et al.  Multiagent systems: a modern approach to distributed artificial intelligence , 1999 .

[11]  Oliver Obst,et al.  Towards a Logical Approach for Soccer Agents Engineering , 2000, RoboCup.

[12]  Peter Baumgartner,et al.  Abductive Coreference by Model Construction , 1999 .

[13]  Manfred Rosendahl,et al.  Specification of Symbols and Implementation of Their Constraints in JKogge , 2000 .

[14]  Joseph Douglas Horton,et al.  Merge Path Improvements for Minimal Model Hyper Tableaux , 1999, TABLEAUX.

[15]  Jürgen Ebert,et al.  A Formalization of SOCCA , 1999 .

[16]  Frieder Stolzenburg,et al.  Loop-Detection in Hyper-Tableaux by Powerful Model Generation , 1999, J. Univers. Comput. Sci..

[17]  Nikos A. Vlassis,et al.  Towards an Optimal Scoring Policy for Simulated Soccer Agents , 2002, RoboCup.