Interleaving Search And Execution
In AI search, there are computational limits and uncertainty due to the opponent’s move as in two-players games. So it is wise to have search and execution interleaved with each search determining only the next move to be made. This paradigm is also applicable to single-agent problems. In the case of autonomous vehicle navigation, for example, information is limited by the horizon of the vehicle, sensors, and it must physically move to acquire more information. Thus, one move must be computed at a time, and that move executed before computing the next. Following are the algorithms designed to address this scenario.