Exchanging and Combining Temporal Information in a Cooperative Environment

Abstract

This paper considers the problem of exchanging and combining temporal information by collaborative agents who act in a dynamic environment. In order to carry out their cooperative activity the agents perform collaborative planning [2] while interleaving planning and execution. In a former paper [3] we presented a mechanism for cooperative planning agents to determine the timetable of the actions that are required to perform their joint activity. In this paper we expand our former work and compare different methods of reasoning and combining temporal information in a team. Determining the time of the actions in a collaborative environment is complex because of the need to coordinate actions of different agents, the partiality of the plans, the partial knowledge on other agents’ activities and on the environment and temporal constraints. Our mechanism focuses on temporal scheduling. Thus, for simplification purposes, the agents do not take into consideration preconditions and effects during their planning process.

This work is supported in part by NSF under Grant No. IIS-9907482. The second author is also affliated with UMIACS.