A theoretical framework for explaining agent behavior
conference paper
To understand emergent processes in multi-agent-based simulations it is important to study the global processes in a simulation as well as the processes on the agent level. The behavior of individual agents is easier to understand when they are able to explain their own behavior. In this paper, a theoretical framework for explaining agent behavior is proposed. By distinguishing different types and contexts of explanations, the framework aims to support the development of explainable agents. The framework is based on an examination of explanation literature, and experiences with developing explainable agents for virtual training. The use of the framework is illustrated by an example about the development of a negotiation agent.
Topics
TNO Identifier
435973
ISBN
9789898425782
Source title
1st International Conference on Simulation and Modeling Methodologies, Technologies and Applications, SIMULTECH 2011, 29 July 2011 through 31 July 2011, Noordwijkerhout
Pages
228-231
Files
To receive the publication files, please send an e-mail request to TNO Repository.