Publications

Automated problem reformulation is a common technique in classical planning to identify and exploit problem structures. Decoupled search is an approach that automatically decomposes planning tasks based on their causal structure, often significantly reducing the search effort. However, its broad applicability is limited by the need for specialized algorithms. In this paper, we present an approach that embodies decoupled search for non-optimal planning through a novel task transformation. Specifically, given a task and a decomposition, we create a transformed task such that the state space of the transformed task is isomorphic to that of decoupled search on the original task. This eliminates the need for specialized algorithms and allows the use of various planning technology in the decoupled-search framework. Empirical evaluation shows that our method is empirically competitive with specialized decoupled algorithms and favorable to other related problem reformulation techniques.

Saturated post-hoc optimization is a powerful method for computing admissible heuristics for optimal classical planning. The approach solves a linear program (LP) for each state encountered during the search, which is computationally demanding. In this paper, we theoretically and empirically analyze to which extent we can reuse an LP solution of one state for another. We introduce a novel sensitivity analysis that can exactly characterize the set of states for which a unique LP solution is optimal. Furthermore, we identify two properties of the underlying LPs that affect reusability. Finally, we introduce an algorithm that optimizes LP solutions to generalize well to other states. Our new algorithms significantly reduce the number of necessary LP computations.

Cost partitioning is the foundation of today’s strongest heuristics for optimal classical planning. However, computing a cost partitioning for each evaluated state is prohibitively expensive in practice. Thus, existing approaches make an approximation and compute a cost partitioning only for a set of sampled states, and then reuse the resulting heuristics for all other states evaluated during the search. In this paper, we present exact methods for cost partitioning heuristics based on linear programming that fully preserve heuristic accuracy while minimizing computational cost. Specifically, we focus on saturated post-hoc optimization and establish several sufficient conditions for when reusing a cost partitioning computed for one state preserves the estimates for other states, mainly based on a sensitivity analysis of the underlying linear program. Our experiments demonstrate that our theoretical results transfer into practice, and that our exact cost partitioning algorithms are competitive with the strongest approximations currently available,while usually requiring fewer linear program evaluations.

Combinatorial reconfiguration is the problem of transforming one solution of a combinatorial problem into another, where each transformation may only apply small changes to a solution and may not leave the solution space. An important example is the independent set reconfiguration (ISR) problem, where an independent set of a graph (a subset of its vertices without edges between them) has to be transformed into another by a sequence of transformations that can replace a vertex in the current subset such that the new subset is still an independent set. The 1st Combinatorial Reconfiguration Challenge (CoRe Challenge 2022) was a competition focused on the ISR problem. The PARIS team successfully participated with two solvers that model the ISR problem as a planning task and employ different planning techniques for solving it. In this work, we describe these models and solvers. For a fair comparison to competing ISR approaches, we re-run the entire competition under equal computational conditions. Besides showcasing the success of planning technology, we hope that this work will create a cross-fertilization of the two research fields.

Matrix multiplication is a fundamental operation of linear algebra, with applications ranging from quantum physics to artificial intelligence. Given its importance, enormous resources have been invested in the search for faster matrix multiplication algorithms. Recently, this search has been cast as a single-player game. By learning how to play this game efficiently, the newly-introduced AlphaTensor reinforcement learning agent discovers many new faster algorithms. In this paper, we show that finding matrix multiplication algorithms can also be cast as a classical planning problem. Based on this observation, we introduce a challenging benchmark suite for classical planning and evaluate state-of-the-art planning techniques on it. We analyze the strengths and limitations of different planning approaches in this domain and show that we can use classical planning to find lower bounds and concrete algorithms for matrix multiplication.

Since its introduction, partial satisfaction planning (PSP), including both oversubscription (OSP) and net-benefit, has received significant attention in the classical planning community. However, hierarchical aspects have been mostly ignored in this context, although several problem domains that form the main motivation for PSP, such as the rover domain, have an inherent hierarchical structure. In this paper, we are taking the necessary steps for facilitating this research direction. First, we formally define hierarchical partial satisfaction planning problems and discuss the usefulness and necessity of this formalism. Second, we present a carefully structured set of benchmarks consisting of OSP and net-benefit problems with hierarchical structure. We describe and analyze the different domains of the benchmark set and the desiderata that are met to provide an interesting and challenging starting point for upcoming research. Third, we introduce various planning techniques that can solve hierarchical OSP problems and investigate their empirical behaviour on our proposed benchmark.

Heuristic search is a successful approach to cost-optimal planning. Bidirectional heuristic search algorithms have been around for a long time, but only recent advances have led to algorithms like BAE* that have the potential to outperform unidirectional heuristic search algorithms like A* in practice. In this work, we analyze BAE* for classical planning and the challenges associated with the underlying assumption of an explicit state representation. We show that it is crucial to use mutexes and reachability analysis to reduce the potentially exponential number of goal states, which makes it possible to create an explicit representation of a reversed planning task that can be used for the backward search of BAE*. Our empirical evaluation shows that BAE* solves more instances than A* in multiple domains with significantly fewer node expansions, demonstrating the usefulness of BAE* in planning.

Cartesian counterexample-guided abstraction refinement (CEGAR) yields strong heuristics for optimal classical planning. CEGAR repeatedly finds counterexamples, i.e., abstract plans that fail for the concrete task. Although there are usually many such abstract plans to choose from, the refinement strategy from previous work is to choose an arbitrary optimal one. In this work, we show that an informed refinement strategy is critical in theory and practice. We demonstrate that it is possible to execute all optimal abstract plans in the concrete task simultaneously, and present methods to minimize the time and number of refinement steps until we find a concrete solution. The resulting algorithm solves more tasks than the previous state of the art for Cartesian CEGAR, both during refinement and when used as a heuristic in an A* search.

In top-k planning, the objective is to determine a set of k cheapest plans that provide several good alternatives to choose from. Such a solution set often contains plans that visit at least one state more than once. Depending on the application, plans with such loops are of little importance because they are dominated by a loopless representative and can prevent more meaningful plans from being found.

In this paper, we motivate and introduce loopless top-k planning. We show how to enhance the state-of-the-art symbolic top-k planner, symK, to obtain an efficient, sound and complete algorithm for loopless top-k planning. An empirical evaluation shows that our proposed approach has a higher k-coverage than a generate-and-test approach that uses an ordinary top-k planner, which we show to be incomplete in the presence of zero-cost loops.

While state-dependent action costs are practically relevant and have been studied before, it is still unclear if they are an essential feature of planning tasks. In this paper, we study to what extent state-dependent action costs are an essential feature by analyzing under which circumstances they can be compiled away. We give a comprehensive classification for combinations of action cost functions and possible cost measures for the compilations.

Our theoretical results show that if both task sizes and plan lengths are to be preserved polynomially, then the boundary between compilability and non-compilability lies between FP and FPSPACE computable action cost functions (under a mild assumption on the polynomial hierarchy). Preserving task sizes polynomially and plan lengths linearly at the same time is impossible.

A key challenge in satisficing planning is to use multiple heuristics within one heuristic search. An aggregation of multiple heuristic estimates, for example by taking the maximum, has the disadvantage that bad estimates of a single heuristic can negatively affect the whole search. Since the performance of a heuristic varies from instance to instance, approaches such as algorithm selection can be successfully applied. In addition, alternating between multiple heuristics during the search makes it possible to use all heuristics equally and improve performance. However, all these approaches ignore the internal search dynamics of a planning system, which can help to select the most useful heuristics for the current expansion step. We show that dynamic algorithm configuration can be used for dynamic heuristic selection which takes into account the internal search dynamics of a planning system. Furthermore, we prove that this approach generalizes over existing approaches and that it can exponentially improve the performance of the heuristic search. To learn dynamic heuristic selection, we propose an approach based on reinforcement learning and show empirically that domain-wise learned policies, which take the internal search dynamics of a planning system into account, can exceed existing approaches.

Cost partitioning admissibly combines the information from multiple heuristics for optimal state-space search. One of the strongest cost partitioning algorithms is saturated cost partitioning. It considers the heuristics in sequence and assigns to each heuristic the minimal fraction of the remaining costs that are needed for preserving all heuristic estimates. Saturated cost partitioning has recently been generalized in two directions: first, by allowing to use different costs for the transitions induced by the same operator, and second, by preserving the heuristic estimates for only a subset of states. In this work, we unify these two generalizations and show that the resulting subset-saturated transition cost partitioning algorithm usually yields stronger heuristics than the two generalizations by themselves.

The objective of optimal oversubscription planning is to find a plan that yields an end state with a maximum utility while keeping plan cost under a certain bound. In practice, the situation occurs whenever a large number of possible, often competing goals of varying value exist, or the resources are not sufficient to achieve all goals. In this paper, we investigate the use of symbolic search for optimal oversubscription planning. Specifically, we show how to apply symbolic forward search to oversubscription planning tasks and prove that our approach is sound, complete and optimal. An empirical analysis shows that our symbolic approach favorably competes with explicit state-space heuristic search, the current state of the art for oversubscription planning.

Symbolic search has proven to be a useful approach to optimal classical planning. In Hierarchical Task Network (HTN) planning, however, there is little work on optimal planning. One reason for this is that in HTN planning, most algorithms are based on heuristic search, and admissible heuristics have to incorporate the structure of the task network in order to be informative. In this paper, we present a novel approach to optimal (totally-ordered) HTN planning, which is based on symbolic search. An empirical analysis shows that our symbolic approach outperforms the current state of the art for optimal totally-ordered HTN planning.

Symbolic search has proven to be a competitive approach to cost-optimal planning, as it compactly represents sets of states by symbolic data structures. While heuristics for symbolic search exist, symbolic bidirectional blind search empirically outperforms its heuristic counterpart and is therefore the dominant search strategy. This prompts the question of why heuristics do not seem to pay off in symbolic search. As a first step in answering this question, we investigate the search behaviour of symbolic heuristic search by means of BDDA*. Previous work identified the partitioning of state sets according to their heuristic values as the main bottleneck. We theoretically and empirically evaluate the search behaviour of BDDA* and reveal another fundamental problem: we prove that the use of a heuristic does not always improve the search performance of BDDA*. In general, even the perfect heuristic can exponentially deteriorate search performance.

The objective of top-k planning is to determine a set of k different plans with lowest cost for a given planning task. In practice, such a set of best plans can be preferred to a single best plan generated by ordinary optimal planners, as it allows the user to choose between different alternatives and thus take into account preferences that may be difficult to model. In this paper we show that, in general, the decision problem version of top-k planning is PSPACE-complete, as is the decision problem version of ordinary classical planning. This does not hold for polynomially bounded plans for which the decision problem turns out to be PP-hard, while the ordinary case is NP-hard. We present a novel approach to top-k planning, called SYM-K, which is based on symbolic search, and prove that SYM-K is sound and complete. Our empirical analysis shows that SYM-K exceeds the current state of the art for both small and large k.

MDPs with factored action spaces, i.e. where actions are described as assignments to a set of action variables, allow reasoning over action variables instead of action states, yet most algorithms only consider a grounded action representation. This includes algorithms that are instantiations of the trial-based heuristic tree search (THTS) framework, such as AO* or UCT.

To be able to reason over factored action spaces, we propose a generalisation of THTS where nodes that branch over all applicable actions are replaced with subtrees that consist of nodes that represent the decision for a single action variable. We show that many THTS algorithms retain their theoretical properties under the generalised framework, and show how to approximate any state-action heuristic to a heuristic for partial action assignments. This allows to guide a UCT variant that is able to create exponentially fewer nodes than the same algorithm that considers ground actions. An empirical evaluation on the benchmark set of the probabilistic track of the latest International Planning Competition validates the benefits of the approach.

Axioms are an extension for classical planning models that allow for modeling complex preconditions and goals exponentially more compactly. Although axioms were introduced in planning more than a decade ago, modern planning techniques rarely support axioms, especially in cost-optimal planning. Symbolic search is a popular and competitive optimal planning technique based on the manipulation of sets of states. In this work, we extend symbolic search algorithms to support axioms natively. We analyze different ways of encoding derived variables and axiom rules to evaluate them in a symbolic representation. We prove that all encodings are sound and complete, and empirically show that the presented approach outperforms the previous state of the art in cost-optimal classical planning with axioms.

The vast majority of work in planning to date has focused on state-independent action costs. However, if a planning task features state-dependent costs, using a cost model with state-independent costs means either introducing a modeling error, or potentially sacrificing compactness of the model. In this paper, we investigate the conflicting priorities of modeling accuracy and compactness empirically, with a particular focus on the extent of the negative impact of reduced modeling accuracy on (a) the quality of the resulting plans, and (b) the search guidance provided by heuristics that are fed with inaccurate cost models. Our empirical results show that the plan suboptimality introduced by ignoring state-dependent costs can range, depending on the domain, from inexistent to several orders of magnitude. Furthermore, our results show that the impact on heuristic guidance additionally depends strongly on the heuristic that is used, the specifics of how exactly the costs are represented, and whether one is interested in heuristic accuracy, node expansions, or overall runtime savings.

Sustainable development is a challenge that needs to be tackled by social consensus. Learning from nature is linked to the hope of learning from biological solutions that have extraordinary qualities. One focus of the publication is to refine the discussion about technology-derived and biology-derived developments by taking descriptive, normative and emotional aspects into consideration. Descriptive aspects are presented on the basis of a straightforward classification tool (decision tree) to clearly describe, distinguish and identify biology-derived and technology-derived developments. A further focus of the article is the presentation of the concept of bioinspired sustainability and the presentation of an evaluation tree for Bioinspired Sustainability Assessment (BiSA).

Symbolic representations have attracted significant attention in optimal planning. Binary Decision Diagrams (BDDs) form the basis for symbolic search algorithms. Closely related are Algebraic Decision Diagrams (ADDs), used to represent heuristic functions. Also, progress was made in dealing with models that take state-dependent action costs into account. Here, costs are represented as Edge-valued Multi-valued Decision Diagrams (EVMDDs), which can be exponentially more compact than the corresponding ADD representation. However, they were not yet considered for symbolic planning.

In this work, we study EVMDD-based symbolic search for optimal planning. We define EVMDD-based representations of state sets and transition relations, and show how to compute the necessary operations required for EVMDD-A*. This EVMDD-based version of symbolic A* generalizes its BDD variant, and allows to solve planning tasks with state-dependent action costs. We prove theoretically that our approach is sound, complete and optimal. Additionally, we present an empirical analysis comparing EVMDD-A* to BDD-A* and explicit A* search. Our results underscore the usefulness of symbolic approaches and the feasibility of dealing with models that go beyond unit costs.

An agent that interacts with a nondeterministic environment can often only partially observe the surroundings. This necessitates observations via sensors rendering more information about the current world state. Sensors can be expensive in many regards therefore it can be essential to minimize the amount of sensors an agent requires to solve given tasks. A limitation for sensor minimization is given by essential sensors which are always required to solve particular problems. In this paper we present an efficient algorithm which determines a set of necessary observation variables. More specifically, we develop a bottom-up algorithm which computes a set of variables which are always necessary to observe, in order to always reach a goal state. Our experimental results show that the knowledge about necessary observation variables can be used to minimize the number of sensors of an agent.