ORTS is a programming environment for studying real-time AI problems such as pathfinding, dealing with imperfect information, scheduling, and planning in the domain of RTS games. These games are fast-paced and very popular. Furthermore, the current state of RTS game AI is bleak which is mainly caused by the lack of planning and learning – areas in which humans are currently much better than machines. Therefore, RTS games make an ideal test-bed for real-time AI research. Unfortunately, commercial RTS games are closed software which prevents researchers from connecting remote AI modules to them. Furthermore, commercial RTS games are based on peer-to-peer technology – which in a nutshell runs the entire simulation on all player machines and just hides part of the game state from the players.
By tampering with the client software it is possible to reveal the entire game state and thereby gain an unfair advantage. We feel that this is unacceptable for playing games on the internet. We therefore started the ORTS project to create a free software system that lets people and machines play fair RTS games. The communication protocol is public and all source code and artwork is freely available. Users can connect whatever client software they like. This is made possible by a server/client architecture in which only the currently visible parts of the game state are sent to the players. This openness leads to new and interesting possibilities ranging from on-line tournaments of autonomous AI players to gauge their playing strength to hybrid systems in which human players use sophisticated GUIs which allow them to delegate tasks to AI helper modules of increasing performance.