Boosting Search Guidance in Problems with Semantic Attachments

Abstract

Most applications of planning to real problems involve complex and often non-linear equations, including matrix operations. PDDL is ill-suited to express such calculations since it only allows basic operations between numeric fluents. To remedy this restriction, a generic PDDL planner can be connected to a specialised advisor, which equips the planner with the ability to carry out sophisticated mathematical operations. Unlike related techniques based on semantic attachment, our planner is able to exploit an approximation of the numeric information calculated by the advisor to compute informative heuristic estimators. Guided by both causal and numeric information, our planning framework outperforms traditional approaches, especially against problems with numeric goals. We provide evidence of the power of our solution by successfully solving four completely different problems.

MPTP: Motion-planning-aware task planning for navigation in belief space

“…In this Section, we validate our approach in two different robot navigation domains, namely office domain and corridor domain as described in Sections 4.1.1 and 4.1.2. We use the temporal POPF-TIF [45] as our task planner by customizing it to achieve semantic attachments of an external module. The external module performs a PRM-based planning in the belief space and is implemented as a dynamically loaded shared library that is passed as an input to the planner.…”

“…The term semantic attachment was coined by Weyhrauch [44] to describe the association of algorithms to function and predicate symbols via external procedures. However, the effects returned by these semantic attachments are not exploited in identifying helpful actions during search and hence do not provide any heuristic guidance, deeming the task unsolvable most often [45]. An action is considered helpful if it achieves at least one of the lowest level goals in the relaxed plan to the state at hand [14].…”

“…An action is considered helpful if it achieves at least one of the lowest level goals in the relaxed plan to the state at hand [14]. Recently, Bernardini et al [45] developed a PDDL-based temporal planner to implicitly trigger such external calls via a specialized semantic attachments called external advisors. They classify variables into direct (V dir ), indirect (V ind ) and free (V free ).…”