Planning in the Cloud: Massively Parallel Planning

This paper describes preliminary work in creating a scalable service for the purposes of automated planning and scheduling: a methodology within artificial intelligence that requires flexible computational resources on a per-problem basis. We describe the challenges of automated planning and how moving solution construction to a distributed system alleviates issues faced in the application of planning in real-world problems. We explore how the current system has been designed and give indication of how this work moves towards creating an online planning service that is scalable to the needs of both individual users and the overall workload required of the system.