The Conflict Partition Scheduling (CPS) is a methodology that builds solutions by repeatedly identifying bottlenecks and posting constraints to solve them. As a result, a temporally flexible schedule is gained, that offers an envelope of executable legal behaviours. CPS has been focused to single agent (e.g. a single rover or spacecraft) scenarios so far, according to a centralized scheduling strategy. The growing interest on space missions composed by coordinated space systems (e.g. spacecraft flotillas and rover teams) makes distributed scheduling technologies even more attractive to develop. The paper extends the CPS approach to a distributed multi-agent scheduling environment. Running isolated and parallel CPSs processes does not guarantee the consistency of inter-agent constraints. The here proposed methodology concerns with controlling the bottleneck conflict partition of each agent through a negotiation mechanism. In particular, each sequencing constraint becomes an agent proposal to negotiate according to a specific strategy. Negotiation mechanism models a majority voting system and reduces the ongoing conflict partitions to propose. The constraint propagations are performed in a Distributed Simple Temporal Network (DisSTN) to keep temporal consistency of the agent society. An analysis on differently sized problems and the results of a two rovers scenario are also offered.
[1]
Nicola Muscettola,et al.
Fast Transformation of Temporal Plans for Efficient Execution
,
1998,
AAAI/IAAI.
[2]
Luke Hunsberger,et al.
Distributing the control of a temporal network among multiple agents
,
2003,
AAMAS '03.
[3]
Nicola Muscettola,et al.
Scheduling by iterative partition of bottleneck conflicts
,
1993,
Proceedings of 9th IEEE Conference on Artificial Intelligence for Applications.
[4]
A. Ercoli Finzi,et al.
Distributed Planning and Scheduling for Space System Flotillas
,
2005
.
[5]
P. Pardalos,et al.
An exact algorithm for the maximum clique problem
,
1990
.
[6]
A. Ercoli Finzi,et al.
A distributed approach for space system formation operation autonomous planning and scheduling
,
2005
.
[7]
P. Pandurang Nayak,et al.
Remote Agent: To Boldly Go Where No AI System Has Gone Before
,
1998,
Artif. Intell..
[8]
Rina Dechter,et al.
Temporal Constraint Networks
,
1989,
Artif. Intell..