PIPSS: Parallel Integrated Planning and Scheduling System

Real problems require the combination of time and resources management. In the last decade, different approaches that integrate planning and scheduling have appeared. In this paper we present PIPSS, a system that combines traditional state space search planning with constraint satisfaction algorithms for scheduling, so that it can also reason about temporal and multicapacity resources problems. PIPSS integrates planning techniques derived from the FF planner and scheduling methods from the OOSCAR scheduler. Its planning and scheduling integration scheme is grounded on a system called IPSS. The most remarkable feature of PIPSS is that it exhibits an open architecture for planning and scheduling integration, based on interfaces, which allows the interoperability of different planning algorithms, scheduling techniques and integration schemes. PIPSS can then be configured to work using any combination of them. Experimental results show that PIPSS outperforms state of the art planner in some of the domains tested.