Intelligent jason agents in virtual soccer simulations

Abstract

The game of virtual soccer requires its players to act as a coherent team, while operating in a highly dynamic environment, with incomplete and unreliable information. As such, the game is often used to test, demonstrate and validate the concepts of multi-agent theory in practice. This paper describes the process of integrating agent-oriented programming language AgentSpeak and its interpreter Jason into SimSpark, the official RoboCup soccer simulator. The end-goal is to design a framework that will enable its users to think about and program soccer playing agents at a more abstract level, in terms of beliefs, goals and plans, allowing for an easier implementation of advanced multi-agent concepts. © 2013 Springer-Verlag Berlin Heidelberg.