Developed by: Juan M. Gálvez, Antonio J. Fernández-Leiva and Carlos Cotta Porras
(University of Malaga, May-2009)
Description
- robotWars is a test environment that
allows two different armies (each of them controlled
by a virtual player - i.e. a game AI) fight in a map (i.e.
the scenario) that contains multiple obstacles and has limited
dimensions.
- It is not a standard game itself in the sense that no human
players intervene interactively during the game; however it
is a perfect scenario to test the goodness and efficacy of
(possibly hand-coded) strategies to control the game AI and
where the human player can influence the game by setting
its initial conditions.
- To provide a good 3D simulation, the
game was programmed with Ogre3D.
Main Components
-
Battle generator: which allows a human to generate different scenarios for
the game by changing the topologies and structures of the
armies; two types of army formation are initially allowed:
testudo or tortoise formation.
-
Battle simulator: which receives as entry two different game AIs
and a map over which a game will be run and provides a
deterministic simulation of a war between the two armies.
Download Links
Referencing the software
The source code of the framework is available to the scientific community under a Creative Commons BY-NC-SA (Attribution + Noncommercial + Sharealike) license. Please cite the following paper when referring to this framework:
-
M. Nogueira, J. Gálvez, C. Cotta, A.J. Fernández Leiva, An Analysis of Hall-of-Fame Strategies in Competitive Coevolutionary Algorithms for Self-Learning in RTS Games. In Learning and Intelligent Optimization Conference 7 (LION 2013), to appear in Lecture Notes in Computer Science, Springer-Verlag, Catania, 2013.