Alfonso Gerevini

Deterministic planning in the fifth international planning competition: PDDL3 and experimental evaluation of the planners

The international planning competition (IPC) is an important driver for planning research. The general goals of the IPC include pushing the state of the art in planning technology by posing new scientific challenges, encouraging direct comparison of planning systems and techniques, developing and improving a common planning domain definition language, and designing new planning domains and problems for the research community. This paper focuses on the deterministic part of the fifth international planning competition (IPC5), presenting the language and benchmark domains that we developed for the competition, as well as a detailed experimental evaluation of the deterministic planners that entered IPC5, which helps to understand the state of the art in the field. We present an extension of pddl, called pddl3, allowing the user to express strong and soft constraints about the structure of the desired plans, as well as strong and soft problem goals. We discuss the expressive power of the new language focusing on the restricted version that was used in IPC5, for which we give some basic results about its compilability into pddl2. Moreover, we study the relative performance of the IPC5 planners in terms of solved problems, CPU time, and plan quality; we analyse their behaviour with respect to the winners of the previous competition; and we evaluate them in terms of their capability of dealing with soft goals and constraints, and of finding good quality plans in general. Overall, the results indicate significant progress in the field, but they also reveal that some important issues remain open and require further research, such as dealing with strong constraints and computing high quality plans in metric-time domains and domains involving soft goals or constraints.

Combining topological and size information for spatial reasoning

Information about the size of spatial regions is often easily accessible and, when combined with other types of spatial information, it can be practically very useful. In this paper we introduce four classes of qualitative and metric size constraints, and we study their integration with the Region Connection Calculus RCC-8, a well-known approach to qualitative spatial reasoning with topological relations. We propose a new path-consistency algorithm for combining RCC-8 relations and qualitative size relations. The algorithm is complete for deciding satisfiability of an input set of topological constraints over one of the three maximal tractable subclasses of RCC-8 containing all the basic relations. Moreover, its time complexity is cubic and is the same as the complexity of the best-known method for deciding satisfiability when only these topological relations are considered. We also provide results on finding a consistent scenario in cubic time for these combined classes. Regarding metric size constraints, we first study their combination with RCC-8 and we show that deciding satisfiability for the combined sets of constraints is NP-hard, even when only the RCC-8 basic relations are used. Then we introduce RCC-7, a subalgebra of RCC-8 that can be used for applications where spatial regions cannot partially overlap. We show that reasoning with the seven RCC-7 basic relations and the universal relation is intractable, but that reasoning with the RCC-7 basic relations combined with metric size information is tractable. Finally, we give a polynomial algorithm for the latter case and a backtracking algorithm for the general case.

An approach to temporal planning and scheduling in domains with predictable exogenous events

The treatment of exogenous events in planning is practically important in many real-world domains where the preconditions of certain plan actions are affected by such events. In this paper we focus on planning in temporal domains with exogenous events that happen at known times, imposing the constraint that certain actions in the plan must be executed during some predefined time windows. When actions have durations, handling such temporal constraints adds an extra difficulty to planning. We propose an approach to planning in these domains which integrates constraint-based temporal reasoning into a graph-based planning framework using local search. Our techniques are implemented in a planner that took part in the 4th International Planning Competition (IPC-4). A statistical analysis of the results of IPC-4 demonstrates the effectiveness of our approach in terms of both CPU-time and plan quality. Additional experiments show the good performance of the temporal reasoning techniques integrated into our planner.

An approach to efficient planning with numerical fluents and multi-criteria plan quality

Dealing with numerical information is practically important in many real-world planning domains where the executability of an action can depend on certain numerical conditions, and the action effects can consume or renew some critical continuous resources, which in pddl can be represented by numerical fluents. When a planning problem involves numerical fluents, the quality of the solutions can be expressed by an objective function that can take different plan quality criteria into account. We propose an incremental approach to automated planning with numerical fluents and multi-criteria objective functions for pddl numerical planning problems. The techniques in this paper significantly extend the framework of planning with action graphs and local search implemented in the lpg planner. We define the numerical action graph (NA-graph) representation for numerical plans and we propose some new local search techniques using this representation, including a heuristic search neighborhood for NA-graphs, a heuristic evaluation function based on relaxed numerical plans, and an incremental method for plan quality optimization based on particular search restarts. Moreover, we analyze our approach through an extensive experimental study aimed at evaluating the importance of some specific techniques for the performance of the approach, and at analyzing its effectiveness in terms of fast computation of a valid plan and quality of the best plan that can be generated within a given CPU-time limit. Overall, the results show that our planner performs quite well compared to other state-of-the-art planners handling numerical fluents.

Generating diverse plans to handle unknown and partially known user preferences

Current work in planning with preferences assumes that user preferences are completely specified, and aims to search for a single solution plan to satisfy these. In many real world planning scenarios, however, the user may provide no knowledge or at best partial knowledge of her preferences with respect to a desired plan. In such situations, rather than presenting a single plan as the solution, the planner must instead provide a set of plans containing one or more plans that are similar to the one that the user really prefers. In this paper, we first propose the usage of different measures to capture the quality of such plan sets. These are domain-independent distance measures based on plan elements (such as actions, states, or causal links) if no knowledge of the user preferences is given, or the Integrated Convex Preference (ICP) measure in case incomplete knowledge of such preferences is provided. We then investigate various heuristic approaches to generate sets of plans in accordance with these measures, and present empirical results that demonstrate the promise of our methods.

Temporal reasoning in Timegraph I–II

We describe two temporal reasoning systems called Timegraph I and II, which can efficiently manage large sets of relations in the Point Algebra as well as metric information. Our representation is based on timegraphs, graphs partitioned into a set of chains on which the search is supported by a metagraph data structure. Timegraph I was originally developed by Taugher, Schubert and Miller in the context of story comprehension. Timegraph II provides useful extensions which makes the system more expressive in the representation of qualitative information and suitable for a larger class of applications. Experimental results show that our approach is very efficient, especially when the given relations admit representation as a collection of chains connected by relatively few cross-chain links.

Combining Topological and Qualitative Size Constraints for Spatial Reasoning

Information about the relative size of spatial regions is often easily accessible and, when combined with other types of spatial information, it can be practically very useful. In this paper we combine a simple framework for reasoning about qualitative size relations with the Region Connection Calculus RCC-8, a widely studied approach for qualitative spatial reasoning with topological relations. Reasoning about RCC-8 relations is NP-hard, but a large maximal tractable subclass of RCC-8 called H8 was identified. Interestingly, any constraint in RCC-8 - H8 can be consistently reduced to a constraint in H8, when an appropriate size constraint between the spatial regions is supplied. We propose an O(n3) time path-consistency algorithm based on a novel technique for combining RCC-8 constraints and relative size constraints, where n is the number of spatial regions. We prove its correctness and completeness for deciding consistency when the input contains topological constraints in H8. We also provide results on finding a consistent scenario in O(n3) time both for combined topological and relative size constraints, and for topological constraints alone. This is an O(n2) improvement over the known methods.

Efficient Plan Adaptation through Replanning Windows and Heuristic Goals

Fast plan adaptation is important in many AI-applications. From a theoretical point of view, in the worst case adapting an existing plan to solve a new problem is no more ecient than a complete regeneration of the plan. However, in practice plan adaptation can be much more ecient than plan generation, especially when the adapted plan can be obtained by performing a limited amount of changes to the original plan. In this paper, we investigate a domain-independent method for plan adaptation that modifies the original plan by replanning within limited temporal windows containing portions of the plan that need to be revised. Each window is associated with a particular replanning subproblem that contains some “heuristic goals” facilitating the plan adaptation, and that can be solved using dierent planning methods. An experimental analysis shows that, in practice, adapting a given plan for solving a new problem using our techniques can be much more ecient than replanning from scratch.

Planning as Propositional CSP: From Walksat to Local Search Techniques for Action Graphs

Graphplan-style of planning can be formulated as an incremental propositional CSP where the (boolean) variables correspond to operator instantiations (actions) that are or are not scheduled at certain time steps. In this paper we present a framework for solving this class of propositional CSPs using local search in planning graphs. The search space consists of particular subgraphs of a planning graph corresponding to (complete) variable assignments, and representing partial plans. The operators for moving from one search state to the next one are graph modifications corresponding to revisions of the current variable assignment (partial plan), or to an extension of the represented CSP.Our techniques are implemented in a planner called LPG using various types of heuristics based on a parametrized objective function, where the parameters weight different constraint violations, and are dynamically evaluated using Lagrange multipliers. LPGs basic heuristic was inspired by Walksat, which in Kautz and Selmans Blackbox can be used to solve the SAT-encoding of a planning graph. An advantage of LPG is that its heuristics exploit the structure of the planning graph, while Blackbox relies on general heuristics for SAT-problems, and requires the translation of the planning graph into propositional clauses. Another major difference is that LPG can handle action execution costs to produce good quality plans. This is achieved by an “anytime” process minimizing an objective function based on the number of constraint violations in a plan and on its overall cost. Experimental results illustrate the efficiency of our approach, showing, in particular, that LPG is significantly faster than Blackbox and other planners based on planning graphs.

On Computing the Minimal Labels in Time Point Algebra Networks

We analyze the problem of computing the minimal labels for a network of temporal relations in point algebra. Van Beek proposes an algorithm for accomplishing this task, which takes O(max(n3, n2 m)) time (for n points and m ≠‐relations). We show that the proof of the correctness of this algorithm given by van Beek and Cohen is faulty, and we provide a new proof showing that the algorithm is indeed correct.