Responsibility-aware Strategic Reasoning in Probabilistic Multi-Agent Systems

Authors: Chunyan Mu, Muhammad Najib, Nir Oren

AAAI 2025 | Venue PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Theoretical We introduce PATL+R, a variant of Probabilistic Alternatingtime Temporal Logic. PATL+R s novelty lies in its incorporation of modalities for causal responsibility, providing a framework for responsibility-aware multi-agent strategic reasoning. We present an approach to synthesise joint strategies that satisfy an outcome specified in PATL+R while optimising the share of expected causal responsibility and reward. This provides a notion of balanced distribution of responsibility and reward gain among agents. To this end, we utilise the Nash equilibrium as the solution concept for our strategic reasoning problem and demonstrate how to compute responsibility-aware Nash equilibrium strategies via a reduction to parametric model checking of concurrent stochastic multi-player games. We show that such a model checking problem can be solved in PSPACE. We further demonstrate how we can utilise the parametric model to compute NE strategies/plans with respect to agents utility functions that consider both reward and responsibility. We show that such a computation can also be done in PSPACE.
Researcher Affiliation Academia 1Department of Computing Science, University of Aberdeen 2Department of Computer Science, Heriot-Watt University
Pseudocode Yes Algorithm 1: Calculate Es, σA(D[CARi,π(ψ)]) Algorithm 2: Calculate Es, σA(D[CPRi,π(φ)])
Open Source Code No The information is insufficient. The paper does not provide any explicit statements about releasing source code or links to a code repository.
Open Datasets No The information is insufficient. The paper uses a conceptual example (Example 1: catching balls) to illustrate the model, but does not use or refer to any publicly available datasets for evaluation.
Dataset Splits No The information is insufficient. The paper does not use external datasets for empirical evaluation, hence no dataset split information is provided.
Hardware Specification No The information is insufficient. The paper focuses on theoretical contributions and formal methods and does not describe the hardware used for any experimental evaluation.
Software Dependencies No The information is insufficient. The paper discusses a new logic and model checking paradigms but does not mention specific software, libraries, or solvers with version numbers.
Experiment Setup No The information is insufficient. The paper focuses on theoretical contributions and formal methods, and does not describe a concrete experimental setup with specific hyperparameters or training configurations.