Basic Algorithm

Here we describe how we put our inference rules together with our representation scheme to efficiently process and glean necessary information about the game state space. The basic algorithm is as follows:

1. At every turn, iterate over all rounds since our last turn, and collect any assertions or clauses about interrogated players that are available in the move history.

2. Apply all of our inference rules on the current game representation.

3. If any inference rules result in any assertions or if any new information is generated, set a dirty flag.

4. If the dirty flag is set, go back to step 2.

5. At this point, no more inferences can be made. Invoke the algorithm for choosing the best interrogation, and then interrogate.

It is very important that we repeat step 3 until no changes occur, since it is not clear in what order the inference rules should be applied so that all possible deductions are made. By iteratively applying the rules, and slowly making progress toward discovering the hidden cards, we ensure that nothing is overlooked.