Robust scheduling and runtime adaptation of multi-agent plan execution

Robustness and reliability with respect to the successful completion of a schedule are crucial requirements for scheduling in multi-agent systems because agent autonomy makes execution environments dynamic and nondeterministic. We introduce a model to incorporate trust which indicates the probability that an agent will comply with its commitments into scheduling, thus improving the predicability and stability of the schedule. To deal with exceptions during execution, we adapt and evolve the schedule at runtime by interleaving the processes of evaluation, scheduling, execution and monitoring in the life cycle of a plan. Experiments show that schedules maximizing participants' trust are more likely to survive and succeed in open and dynamic environments. The results also prove that the proposed plan evaluation approach conforms with the simulation result, thus being helpful for plan selection.