Computing Programs for Generalized Planning as Heuristic Search

TL;DR

This paper introduces a novel heuristic search method for Generalized Planning, defining a program-based solution space and the BFGP algorithm to efficiently solve GP problems regardless of instance size.

Contribution

It adapts the heuristic search paradigm to Generalized Planning by defining a program-based solution space and proposing the BFGP algorithm for effective search.

Findings

First native heuristic search approach to GP

Program-based solution space independent of instance size

BFGP algorithm guided by evaluation and heuristic functions

Abstract

Although heuristic search is one of the most successful approaches to classical planning, this planning paradigm does not apply straightforwardly to Generalized Planning (GP). This paper adapts the planning as heuristic search paradigm to the particularities of GP, and presents the first native heuristic search approach to GP. First, the paper defines a program-based solution space for GP that is independent of the number of planning instances in a GP problem, and the size of these instances. Second, the paper defines the BFGP algorithm for GP, that implements a best-first search in our program-based solution space, and that is guided by different evaluation and heuristic functions.