Meirav Hadad and Sarit Kraus
We present strategies for reasoning and combining temporal information of intelligent agents. The agents situated in a dynamic environment collaborate with other agents and interleave planning and acting. Our work, which is based on a formal model of collaborative planning, allows each agent to develop its timetable individually and to take into consideration various types of time constraints of its activities and of its collaborators. However, its individual timetable can be easily integrated with the activities of other agents. Thus, in contrast to other works on multi-agent systems, which suggest that either the team members maintain full synchronization by broadcasting messages or that a team leader determines the times of actions, our mechanism enables the individual agent to determine its time frame individually and to be integrated easily in the activities of other agents. We have developed a simulation of such environments and present simulation results comparing different methods of exchanging temporal information among team members.