Knowledge based programs: on the complexity of perfect recall in finite environments

Authors:
Ron van der Meyden
Affiliations:
University of Technology, Sydney, Broadway NSW, Australia
Venue:
TARK '96 Proceedings of the 6th conference on Theoretical aspects of rationality and knowledge
Year:
1996

Citing 11
Cited 0

How processes learn

Distributed Computing
A knowledge-theoretic analysis of atomic commitment protocols

PODS '87 Proceedings of the sixth ACM SIGACT-SIGMOD-SIGART symposium on Principles of database systems
Knowledge and common knowledge in a distributed environment

Journal of the ACM (JACM)
Knowledge and common knowledge in a byzantine environment: crash failures

Information and Computation
A little knowledge goes a long way: knowledge-based derivations and correctness proofs for a family of protocols

Journal of the ACM (JACM)
Reasoning about knowledge

Reasoning about knowledge
Knowledge-based programs

Proceedings of the fourteenth annual ACM symposium on Principles of distributed computing
Alternation

Journal of the ACM (JACM)
A Knowledge Theoretic Account of Recovery in Distributed Systems: The Case of Negotiated Commitment

Proceedings of the 2nd Conference on Theoretical Aspects of Reasoning about Knowledge
Implementing Knowledge-Based Programs

Proceedings of the Sixth Conference on Theoretical Aspects of Rationality and Knowledge
Common knowledge and update in finite environments. I: extended abstract

TARK '94 Proceedings of the 5th conference on Theoretical aspects of reasoning about knowledge

Quantified Score

Hi-index	0.00

Visualization

Abstract

Knowledge based programs have been proposed as an abstract formalism for the design of multi-agent protocols, based on the idea that an agent's actions are a function of its state of knowledge. The key questions in this approach concern the relationship between knowledge based programs and their concrete implementations. We present a variant of the framework of Fagin et al. that facilitates the study of a certain sort of optimization of these implementations. Within this framework, we investigate the inherent complexity of the implementations of atemporal knowledge based programs under the assumptions that the environment is finite state, and that agents operate synchronously and with perfect recall. We provide a simple example showing that one cannot expect to always obtain finite state implementations under this assumption. In fact, we show there exist environments in which knowledge based programs may generate behaviour of PSPACE-complete complexity. This is the most complex behaviour possible given our assumptions.