Computing Parameter Domains as an Aid to Planning

Weshowthat by inferring parametexdomainsof planningoperators,given the definitions of the operators and the initial and goal conditions, we can often speed up the planning process. We infer parameter domains by a polynomial-time algorithm that uses forward propagation of sets of constants occurring in the initial conditions and in operator postconditions. Duringplanning parameter domains can be used to prune operator instances whoseparameter domainsare inconsistent with bindingconstraints, and to eliminate spurious "clobbering threats" that cannot, in fact, be realized withoutviolating domainconstraints. Weillustrate these applications withexamples from the UCPOPtest suite and from the Rochester TRAINStransportation planning domain.

• Type: Conference
• Year: 1996
• Where: AIPS
• Authors: Alfonso Gerevini, Lenhart K. Schubert
• Keywords: AIPS 1996, Artificial Intelligence, Initial Conditions, Operator Postconditions, Parameter Domains