Reasoning about Strategic Abilities: Agents with Truly Perfect Recall.

In alternating-time temporal logic ATL*, agents with perfect recall assign choices to sequences of states, i.e., to possible finite histories of the game. However, when a nested strategic modality is interpreted, the new strategy does not take into account the previous sequence of events. It is as if agents collect their observations in the nested game again from scratch, thus, effectively forgetting what they observed before. Intuitively, it does not fit the assumption of agents having perfect recall of the past. In this article, we investigate the alternative semantics for ATL* where the past is not forgotten in nested games. We show that the standard semantics of ATL* coincides with the “truly perfect recall” semantics for agents with perfect information and in case of so-called “objective” abilities under uncertainty. On the other hand, the two semantics differ significantly for the most popular (“subjective”) notion of ability under imperfect information. The same applies to the standard vs. “truly perfect recall” semantics of ATL* with persistent strategies. We compare the relevant variants of ATL* by looking at their expressive power, sets of validities, and tractability of model checking.