Implementing Reversible Processes in Multi-agent Action Languages Using Answer Set Planning

摘要

This paper presents an implementation of the action language L~(mt) in answer set programming. The novelty of this language comes from the use of processes to execute delayed effects for actions. In addition, the ability to reverse, or cancel, the processes is available. A simple example is introduced to show when reversing actions are useable - even by other agents in the domain. These processes are the base foundation for future implementation of commitments in planning for multi-agent domains.