An Implementation of a Multi-Agent Plan Synchronizer

A program is described which augments plans with synchronizing primitives to ensure appropriate conflict avoidance and cooperation The plans are particularly suitable for describing the activity of multiple agents which may interfere with each other The interpretation of a plan is given as a non deterministic finite automaton which exchanges messages with an environment for the commencement and conclusion of primitive actions which take place over a period of time The synchronized plan allows any and all execution sequences of the original plan which guarantee correct interaction 1. Introduction All planning systems operate by combining actions or sub-plans in some way so that the total plan satisfies some constraint usually to achieve a goal state. Knowledge of how actions interact with each other and the world is used to determine the appropriate combinations In NOAH[5], for example, consideration of interactions between sub-plans is explicit in the planning process A plan is a partial order of sub-plans The planning technique is to expand sub-plans into partial orderings of lower level sub-plans, and then look for and resolve ensuing conflicts The resolution may impose fixed orderings Thi s can be unnecessarily restrictive, as in the case where two sub-plans may execute in any order but not at the same time. This paper considers the use of synchronizing primitives to resolve conflicts and produce a plan which is as unrestrictive as possible The model s of plan and action used are appropriate for simple agents or robots engaged in parallel and or repetitive tasks which may be described at a level not using sensory input to the agents. Use could be made of this technique, for example, in automated assembl y lines Georgeff [2] has also done related work on planning for multiple agents