Generating Cyclic Plans for Traffic Signals

In the AI literature, plan generation assumes non-cyclic plans, i.e. the initial state does not match the goal state. However, in the real world there are domains in which plans may be cyclic, possibly repeating. Examples include the hiking, fly -drive, and mountain climbing domains. Another is developing traffic signals plans, i.e. traffic -light sequences. The purpose of this paper is to describe the traffic signals engineering domain, focusing on the existing process for generating cyclic plans, and identifying possible approaches for automating it. The next step in this work in progress is to select an approach for developing a prototype.