Combining Learning Techniques for Classical Planning: Macro-operators and Entanglements

Abstract—Planning techniques recorded a significant progress
during recent years. However, many planning problems remain
still hard and challenging. One of the most promising approaches
is gathering additional knowledge by using learning techniques.
Advantageously, many sorts of knowledge can be encoded back
into planning domains (or problems) and common planning
systems can be applied on them. Macro-operators are well known
sort of knowledge. Macro-operators are operators that represent
a sequence of primitive planning operators that are formalized
like ‘normal‘ planning operators. The other sort of knowledge
consists of pruning unnecessary operators’ instances (actions) by
investigating connections (entanglements) between operators and
initial or goal predicates. In this paper, we will show how we can
put these approaches together. We will of course experimentally
evaluate how the performance of planners is improved. The
experiments showed that combining of these learning techniques
can improve the planning process.