Seminars

A Logic for Strategic Reasoning of Rational Agents with Normative Constraints

Date15 Apr 2024
Time12:30
LocationOnline
SpeakerValentin Goranko

Recording: Watch on YouTube

The next talk of our seminar series will be given by

Valentin Goranko

(Stockholm University) via Teams (please, see details below) on Monday,

April 15th

, starting from

12:30

CET. Here it is the title and abstract of his talk:

Title:

A Logic for Strategic Reasoning of Rational Agents with Normative Constraints

Abstract:

I will present a logic-based study of agency and multi-agent systems which brings together rationality and normativity. In particular, I will introduce, discuss, and illustrate with some examples the core version of the Logic for Strategic Reasoning with Normative Constraints SRNC, which is a formal logical system for reasoning about the strategic abilities of rational agents which act within a multi-agent system in pursuit of explicitly specified goals, in the context of individual and collective norms, such as permissions, obligations and prohibitions, which constrain accordingly their behaviour. A typical such normative constraint is of the type:

“The agent a may / must / may not perform an action satisfying a given condition C.”

These constraints naturally generalise to groups (coalitions) of agents.

The core version of the logic SRNC which I will present extends the Coalition Logic CL by conditioning the coalitional strategic operators of CL with such normative constraints, intuitively stated as follows:

“The coalition A has a joint action/strategy satisfying the system of normative constraints N(A) which guarantees achievement of a given goal G against any actions/strategies taken by the agents not in A which satisfy the normative constraints in N(A).”

Teams Link

:

https://teams.microsoft.com/l/meetup-join/19%3ameeting_ZWI2ODQyMzAtMDUwZi00ZDFmLTk1MGMtZDBlNjdmNGE3YjQ4%40thread.v2/0?context=%7b%22Tid%22%3a%2213b55eef-7018-4674-a3d7-cc0db06d545c%22%2c%22Oid%22%3a%2209cd4e85-c5dc-4942-8756-2b1388a51272%22%7d

Anyone who is interested is welcome to join.