Ioannis Refanidis, Ioannis Vlahavas
In this paper we present GRT-R, an enhanced version of the GRT planner with the ability to explicitly handle resources. GRT is a domain independent heuristic planner for STRIPS worlds, which work in the space of the states. The heuristic computes off-line, in a pre-processing phase, estimates for the distances between the domain's facts and the goals. These estimates are utilized during the search process, in order to obtain values for the distances between the intermediate states and the goals. We propose to explicit represent resources in a numerical way. We consider resources that are only consumed and they cannot be produced either by internal or external actions. GRT-R uses vectors of values, where the first one estimates the distance between a fact and the goals, while the remaining values estimate the amount of resources needed to achieve that fact. GRT-R assigns each fact with a number of such vectors, each one of them having minimum value to either at least one resource or the distance. These vectors correspond to alternative ways to achieve the fact. Concerning the search process, GRT-R assigns similarly each intermediate state with a vector of values. This vector is based on the vectors of the state's facts, with the intention minimize the distance between the state and the goals, without to exceed the available resources. Performance results show that GRT-R copes well in domains that embody resrources, like the Mystery introduced in the AIPS-98 planning competition.
Keywords: Planning, Search, Automated Reasoning
Citation: Ioannis Refanidis, Ioannis Vlahavas: Heuristic Planning with Resources. In W.Horn (ed.): ECAI2000, Proceedings of the 14th European Conference on Artificial Intelligence, IOS Press, Amsterdam, 2000, pp.521-525.