Engineering Benchmarks for Planning: the Domains Used in the  Deterministic Part of IPC-4

Hoffmann, Jörg; Edelkamp, Stefan; Thiébaux, Sylvie; Englert, Roman; Liporace, Frederico dos Santos; Trüg, Sebastian

doi:10.1613/jair.1982

Cited by 41 publications

(35 citation statements)

References 48 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…AI planning has recently made great advances. The evolution of the international planning competition over its four editions (Bacchus, 2001;Hoffmann, Edelkamp, Englert, Liporace, Thiébaux, & Trüg, 2004;Long & Fox, 2003;McDermott, 2000) accurately reflects this. Successive editions introduced more and more complex and realistic benchmarks, or harder problem instances in the same domain.…”

Section: Introductionmentioning

confidence: 96%

Macro-FF: Improving AI Planning with Automatically Learned Macro-Operators

Botea¹,

Enzenberger²,

Müller³

et al. 2005

jair

121

145

View full text Add to dashboard Cite

Despite recent progress in AI planning, many benchmarks remain challenging for current planners. In many domains, the performance of a planner can greatly be improved by discovering and exploiting information about the domain structure that is not explicitly encoded in the initial PDDL formulation. In this paper we present and compare two automated methods that learn relevant information from previous experience in a domain and use it to solve new problem instances. Our methods share a common four-step strategy. First, a domain is analyzed and structural information is extracted, then macro-operators are generated based on the previously discovered structure. A filtering and ranking procedure selects the most useful macro-operators. Finally, the selected macros are used to speed up future searches.We have successfully used such an approach in the fourth international planning competition IPC-4. Our system, Macro-FF, extends Hoffmann's state-of-the-art planner FF 2.3 with support for two kinds of macro-operators, and with engineering enhancements. We demonstrate the effectiveness of our ideas on benchmarks from international planning competitions. Our results indicate a large reduction in search effort in those complex domains where structural information can be inferred.

show abstract

Section: Introductionmentioning

confidence: 96%

Macro-FF: Improving AI Planning with Automatically Learned Macro-Operators

Botea¹,

Enzenberger²,

Müller³

et al. 2005

jair

121

145

View full text Add to dashboard Cite

show abstract

“…Line switching is also used to reconfigure a network to minimise line losses or balance load. The PSR problem appeared as a benchmark of the 2004 International Planning Competition (Hoffmann, Edelkamp, Thiébaux, Englert, dos S. Liporace, & Trüg, 2006). In contrast to the IPC benchmark, our formalism allows us to model numeric nonlinear power flows as well as capacity and voltage constraints, which is an essential requirement to make the model realistic.…”

Section: Switching Problems In Power Networkmentioning

confidence: 99%

Extending Classical Planning with State Constraints: Heuristics and Search for Optimal Planning

Haslum¹,

Ivankovic²,

Ramírez³

et al. 2018

jair

View full text Add to dashboard Cite

We present a principled way of extending a classical AI planning formalism with systems of state constraints, which relate -sometimes determine -the values of variables in each state traversed by the plan. This extension occupies an attractive middle ground between expressivity and complexity. It enables modelling a new range of problems, as well as formulating more efficient models of classical planning problems. An example of the former is planning-based control of networked physical systems -power networks, for examplein which a local, discrete control action can have global effects on continuous quantities, such as altering flows across the entire network. At the same time, our extension remains decidable as long as the satisfiability of sets of state constraints is decidable, including in the presence of numeric state variables, and we demonstrate that effective techniques for costoptimal planning known in the classical setting -in particular, relaxation-based admissible heuristics -can be adapted to the extended formalism. In this paper, we apply our approach to constraints in the form of linear or non-linear equations over numeric state variables, but the approach is independent of the type of state constraints, as long as there exists a procedure that decides their consistency. The planner and the constraint solver interact through a well-defined, narrow interface, in which the solver requires no specialisation to the planning context. Furthermore, we present an admissible search algorithm -a variant of A -that is able to make use of additional information provided by the search heuristic, in the form of preferred actions. Although preferred actions have been widely used in satisficing planning, we are not aware of any previous use of them in optimal planning.

show abstract

“…For the Satellite domain, for instance, we can adopt the respective aspects from the IPC-4 benchmark suite [138]. The UM Translog enhancements can be inspired by the appropriate IPC definition [290] and the corresponding demonstration scenarios that have been translated for the SHOP2 system.…”

Section: Domain Model Specificationsmentioning

confidence: 99%

Hybrid Planning and Scheduling

Schattenberg¹

2015

Künstl Intell

View full text Add to dashboard Cite

PrefacePlanning and scheduling are well-established disciplines in the field of Artificial Intelligence. They provide flexibility, robustness, and effectiveness to complex software systems in a variety of application areas. While planning is the process of finding a course of action that achieves a goal or performs a specified task, scheduling deals with the assignment of resources and time to given activities, taking into account resource restrictions and temporal dependencies. In other words, planning focuses on reasoning about causal structures and identifying the necessary actions for achieving a specific goal; scheduling concentrates on resource consumption and production for optimizing a coherent parameter assignment of a plan. As successful these techniques clearly are, the actual demands of complex, real-world applications go far beyond the potential of these single methods, however. They require an adequate integration of these problem solving methods as well as a combination of different planning and scheduling paradigms. Particularly important are abstraction-based, hierarchical approaches because of both their expressive knowledge representation and their efficiency in finding solutions. Current state-of-the-art systems rarely address the question of method integration; isolated approaches do so only in ad hoc implementations and mostly lack a proper formal basis.This thesis presents a formal framework for plan and schedule generation based on a well-founded conceptualization of refinement planning: An abstract problem specification is transformed stepwise into a concrete, executable solution. In each refinement step, plan deficiencies identify faulty or under-developed parts of the plan, which in turn triggers the generation of transformation operators that try to resolve them. All involved entities are explicitly represented and therefore transparent to the framework. This property allows for two novel aspects of our approach: First, any planning and scheduling methodology can be functionally decomposed and mapped on the deficiency announcement and plan transformation generation principle, and second, the framework allows for an explicit declaration of planning strategies. We first investigate the flexibility of the extremely modular system design by instantiating the framework in a variety of system configurations including classical partial-order causal-link (POCL) planning, hierarchical task-network (HTN) planning, and classical scheduling.As a key feature, the presented approach provides a formally integrated treatment of action and state abstraction, thus naturally combining causality-focused reasoning with hierarchical, procedure-oriented methods. While the use of procedural knowledge allows to rely on well-known, predefined solutions to planning problems, the non-hierarchical methods provide the flexibility to come up with non-standard solutions and to complete under-specified problem instances, respectively. The resulting technique of hybrid planning is capable of constructing a plan's caus...

show abstract

Engineering Benchmarks for Planning: the Domains Used in the Deterministic Part of IPC-4

Cited by 41 publications

References 48 publications

Macro-FF: Improving AI Planning with Automatically Learned Macro-Operators

Macro-FF: Improving AI Planning with Automatically Learned Macro-Operators

Extending Classical Planning with State Constraints: Heuristics and Search for Optimal Planning

Hybrid Planning and Scheduling

Contact Info

Product

Resources

About