Combining Domain-Independent Planning and HTN Planning: The Duet Planner

  title={Combining Domain-Independent Planning and HTN Planning: The Duet Planner},
  author={Alfonso Gerevini and Ugur Kuter and Dana S. Nau and Alessandro Saetti and Nathaniel Waisbrot},
Despite the recent advances in planning for classical domains, the question of how to use domain knowledge in planning is yet to be completely and clearly answered. Some of the existing planners use domain-independent search heuristics, and some others depend on intensively-engineered domain-specific knowledge to guide the planning process. In this paper, we describe an approach to combine ideas from both of the above schools of thought. We present Duet, our planning system that incorporates… 

Figures and Tables from this paper

A Translation-based Approach to Hierarchical Task Network Planning
This paper presents an effective way to translate a HTN planning domain specification into the corresponding standard PDDL specification, to allow even many domain-independent classical planners to utilize domain-specific control knowledge contained in the HTN specifications.
The GoDeL Planning System: A More Perfect Union of Domain-Independent and Hierarchical Planning
A simple formalism that extends classical planning to include problem decomposition using methods, and a planning algorithm based on this formalism, GoDeL (Goal Decomposition with Landmarks), is defined, which is sound and complete irrespective of whether the domain knowledge is complete.
On the Use of Landmarks in LPG
This work proposes the use of Landmarks for the LPG planner, considering different design choices and analysing empirically its impact on the performance of the planner, finding results comparable with the state-of-the-art planner LAMA.
A hierarchical goal-based formalism and algorithm for single-agent planning
A hierarchical goal-based planning formalism and a planning algorithm, GDP (Goal-Decomposition Planner), that combines some aspects of both HTN planning and domain-independent planning and allows the planning agent to use domain- independent heuristic functions to guide the application of both methods and actions.
A Generic Method to Guide HTN Progression Search with Classical Heuristics
This paper provides an improved progression algorithm, proves its correctness, and empirically shows its efficiency gain, and presents an approach that allows to use arbitrary classical (non-hierarchical) heuristics in HTN planning.
HTN Planning as Heuristic Progression Search
This article introduces two novel progression algorithms that avoid unnecessary branching when the problem at hand is partially ordered and shows that both are sound and complete and introduces a method to apply arbitrary classical planning heuristics to guide the search in HTN planning.
An Admissible HTN Planning Heuristic
It is shown that rebuilding the so-called task decomposition graph (TDG) during planning can improve heuristic accuracy thereby decreasing the explored search space, and an admissible heuristic for standard HTN planning is proposed.
Translating HTNs to PDDL: A Small Amount of Domain Knowledge Can Go a Long Way
It is shown how to translate HTN domain descriptions into PDDL so that they can be used by classical planners, and that even small and incomplete amounts of HTN knowledge can greatly improve a classical planner's performance.
A Unified Framework for Knowledge-Lean and Knowledge-Rich Planning
A unified framework that supports both approaches to planning, along with an implemented system that embodies its tenets is presented, and it is demonstrated that the system can handle problems stated in terms of goals, tasks, or their combination.


Using temporal logics to express search control knowledge for planning
Planning Through Stochastic Local Search and Temporal Action Graphs in LPG
This paper focuses on temporal planning, introducing TA-graphs and proposing some techniques to guide the search in LPG using this representation, and shows that these techniques can be very effective.
Macro-FF: Improving AI Planning with Automatically Learned Macro-Operators
Two automated methods that learn relevant information from previous experience in a domain and use it to solve new problem instances are presented and compared and indicate a large reduction in search effort in those complex domains where structural information can be inferred.
Towards Model-lite Planning : A Proposal For Learning & Planning with Incomplete Domain Models
This paper describes a new learning and planning scheme for the model-lite planning paradigm, which motivates the study of planning techniques that do not rely on the complete domain definition, and uses probabilistic logic to represent uncertain domain models.
Hybrid Planning for Partially Hierarchical Domains
This paper provides a generalized plan-space refinement that is capable of handling non-primitive actions and provides a principled way of handling partially hierarchical domains, while preserving systematicity, and respecting the user-intent inherent in the reduction schemas.
O-Plan: The open Planning Architecture
Model-lite Planning for the Web Age Masses: The Challenges of Planning with Incomplete and Evolving Domain Models
This work motivates model-lite planning technology aimed at reducing the domain-modeling burden (possibly at the expense of reduced functionality), and outlines the research challenges that need to be addressed to realize it.
Temporal Planning using Subgoal Partitioning and Resolution in SGPlan
This paper presents a partition-and-resolve strategy that looks for locally optimal subplans in constraint-partitioned temporal planning subpro problems and that resolves those inconsistent global constraints across the subproblems.
Planning as Heuristic Search: New Results
Hspr is a heuristic search planner that searches backward from the goal rather than forward from the initial state, which allows hspr to compute the heuristic estimates only once, and can produce better plans, often in less time.
The Fast Downward Planning System
  • M. Helmert
  • Computer Science
    J. Artif. Intell. Res.
  • 2006
A full account of Fast Downward's approach to solving multivalued planning tasks is given and a new non-heuristic search algorithm called focused iterative-broadening search, which utilizes the information encoded in causal graphs in a novel way is presented.