Speaker: Valentin Goranko (Stockholm University)
Date and Time: Wednesday, April 17th 2019, 16:30-18:00
Venue: ILLC Seminar Room F1.15, Science Park 107.
Title: Towards a Logic for Conditional Strategic Reasoning.
Abstract. This talk is based on a work in progress, joint with Fengkui Ju, Beijing Normal University.
We consider systems of rational agents who act in pursuit of their individual and collective objectives and we study the reasoning of an agent or an external observer about the consequences from the expected choices of action of the other agents, based on their objectives, in order to assess the agent’s ability to achieve her own objective.
For instance, consider a scenario where an agent, Alice, has an objective O(A) to achieve. Suppose, that Alice has several possible choices of an action (or, strategy) that would possibly, or certainly, guarantee the achievement of her objective O(A). Now, Bob, another agent or an external observer, is reasoning about the consequences from Alice’s possible actions with respect to the occurrence of Bob’s objective or intended outcome O(B). Depending on Bob’s knowledge about Alice’s objective and of her available strategic choices that would guarantee the achievement of that objective, there can be several possible cases for Bob’s reasoning, based on whether or not Bob knows Alice’s objective, her possible actions towards achieving that objective, and her intentions on how to act. Thus, Bob has to reason about his own abilities to achieve his objective O(B), conditional on what he knows or expect that Alice may decide to do. That scenario naturally extends to several agents reasoning about their abilities conditional on how they expect the others to act.
In some of these cases, reasoning about conditional strategic abilities can be formally described in Coalition Logic (CL) or in its temporal extension, the alternating time temporal logic ATL(*), with semantics involving strategy contexts. Other cases, however, require new and more expressive operators capturing more refined patterns of strategic ability, especially suited for the kind of conditional strategic reasoning in scenarios described above. In our work we introduce and study such new operators and logical languages that can capture variations of conditional strategic reasoning.
In this talk, after an informal discussion of conditional strategic reasoning, I will introduce some new strategic operators capturing such reasoning, will provide formal semantics for the resulting logical languages and will discuss briefly their expressiveness and some of their meta-properties. I will conclude with some open questions for future research.