IPSS: A Hybrid Reasoner for Planning and Scheduling

In this paper we describe IPSS (Integrated Planning and Scheduling System), a domain independent solver that integrates an AI heuristic planner, that synthesizes courses of actions, with a constraint-based scheduler for reasoning about time and resources. IPSS is able to solve planning problems with time (precedence constraints, deadline, time windows, etc) and binary resource usage/consumption. Experimental results show that the contextual reasoning of the planner with the constraint-based solver allows to improve the total makespan on a set of problems characterized by multiple agents. 1 IPSS: AN INTEGRATED SYSTEM IPSS is a hybrid software architecture with two parallel processes (planning and scheduling) “collaborating” to solve the same problem relying on different representations. It is mainly subdivided in two systems (see Figure 1): IPSS-P that corresponds to the planning reasoner and IPSS-S that corresponds to the scheduler. On one hand, the planner focuses on the actions s...

• Type: Conference
• Year: 2004
• Where: ECAI
• Authors: María Dolores Rodríguez-Moreno, Angelo Oddi, Daniel Borrajo, Amedeo Cesta, Daniel Meziat
• Keywords: AI Heuristic Planner, Artificial Intelligence, Constraint-based Scheduler, Domain Independent Solver, ECAI 2004