Label reduction is a technique for simplifying families of labeled transition systems by dropping distinctions between certain transition labels. While label reduction is critical to the efficient computation of merge-and-shrink heuristics, current theory only permits reducing labels in a limited number of cases. We generalize this theory so that labels can be reduced in every intermediate abstraction of a merge-and-shrink tree. This is particularly important for efficiently computing merge-and-shrink abstractions based on non-linear merge strategies. As a case study, we implement a non-linear merge strategy based on the original work on merge-and-shrink heuristics in model checking by Dräger et al.

There are two complementary ways to evaluate planning algorithms: performance on benchmark problems derived from real applications and analysis of performance on parametrized families of problems with known properties. Prior to this work, few means of generating parametrized families of hard planning problems were known. We generate hard planning problems from the solvable/unsolvable phase transition region of well-studied NP-complete problems that map naturally to navigation and scheduling, aspects common to many planning domains. We observe significant differences between state-of-the-art planners on these problem families, enabling us to gain insight into the relative strengths and weaknesses of these planners. Our results confirm exponential scaling of hardness with problem size, even at very small problem sizes. These families provide complementary test sets exhibiting properties not found in existing benchmarks.

Utilizing multiple queues in Greedy Best-First Search (GBFS) has been proven to be a very effective approach to satisficing planning. Successful techniques include extra queues based on Helpful Actions (or Preferred Operators), as well as using Multiple Heuristics. One weakness of all standard GBFS algorithms is their lack of exploration. All queues used in these methods work as priority queues sorted by heuristic values. Therefore, misleading heuristics, especially early in the search process, can cause the search to become ineffective. Type systems, as introduced for heuristic search by Lelis et al, are a development of ideas for exploration related to the classic stratified sampling approach. The current work introduces a search algorithm that utilizes type systems in a new way – for exploration within a GBFS multiqueue framework in satisficing planning. A careful case study shows the benefits of such exploration for overcoming deficiencies of the heuristic. The proposed new baseline algorithm Type-GBFS solves almost 200 more problems than baseline GBFS over all International Planning Competition problems. Type-LAMA, a new planner which integrates Type-GBFS into LAMA-2011, solves 36.8 more problems than LAMA-2011.

Backdoors measure the distance to tractable fragments and have become an important tool to find fixed-parameter tractable (fpt) algorithms. Despite their success, backdoors have not been used for planning, a central problem in AI that has a high computational complexity. In this work, we introduce two notions of backdoors building upon the causal graph. We analyze the complexity of finding a small backdoor (detection) and using the backdoor to solve the problem (evaluation) in the light of planning with (un)bounded plan length/domain of the variables. For each setting we present either an fpt-result or rule out the existence thereof by showing parameterized intractability. In three cases we achieve the most desirable outcome: detection and evaluation are fpt.

A wide range of robotic missions contain actions that exhibit looping behavior. Examples of these actions include picking fruit in agriculture, pick-and-place tasks in manufacturing and search patterns in robotic search or survey missions. These looping actions often have a range of acceptable values for the number of loops and a preference function over them. For example, during robotic survey missions, the information gain is expected to increase with the number of loops in a search pattern. Since these looping actions also take time, which is typically bounded, there is a challenge of maximizing utility while respecting time constraints. In this paper, we introduce the Looping Temporal Problem with Preference (LTPP) as a simple parameterized extension of a simple temporal problem. In addition, we introduce a scheduling algorithm for LTPPs which leverages the structure of the problem to find the optimal solution efficiently. We show more than an order of magnitude improvement in run-time over current scheduling techniques and framing a LTPP as a MINLP.

Planning with sensing actions under partial observability is a computationally challenging problem that is fundamental to the realization of AI tasks in areas as diverse as robotics, game playing, and diagnostic problem solving. Recent work on generating plans for partially observable domains has advocated for online planning, claiming that offline plans are often too large to generate. Here we push the envelope on this challenging problem, proposing a technique for generating conditional (aka contingent) plans offline. The key to our planner's success is the reliance on state-of-the-art techniques for fully observable non-deterministic (FOND) planning. In particular, we use an existing compilation for converting a planning problem under partial observability and sensing to a FOND planning problem. With a modified FOND planner in hand, we are able to scale beyond previous techniques for generating conditional plans with solutions that are orders of magnitude smaller than previously possible in some domains.

Monte Carlo tree search (MCTS) algorithms are a popular approach to online decision-making in Markov decision processes (MDPs). These algorithms can, however, perform poorly in MDPs with high stochastic branching factors. In this paper, we study state aggregation as a way of reducing stochastic branching in tree search. Prior work has studied formal properties of MDP state aggregation in the context of dynamic programming and reinforcement learning, but little attention has been paid to state aggregation in MCTS. Our main result is a performance loss bound for a class of value function-based state aggregation criteria in expectimax search trees. We also consider how to construct MCTS algorithms that operate in the abstract state space but require a simulator of the ground dynamics only. We find that trajectory sampling algorithms like UCT can be adapted easily, but that sparse sampling algorithms present difficulties. As a proof of concept, we experimentally confirm that state aggregation can improve the finite-sample performance of UCT.

Planning in hybrid domains is an important and challenging task, and various planning algorithms have been proposed in the last years. From an abstract point of view, hybrid planning domains are based on hybrid automata, which have been studied intensively in the model checking community. In particular, powerful model checking algorithms and tools have emerged for this formalism. However, despite the quest for more scalable planning approaches, model checking algorithms have not been applied to planning in hybrid domains so far. In this paper, we make a first step in bridging the gap between these two worlds. We provide a formal translation scheme from PDDL+ to the standard formalism of hybrid automata, as a solid basis for using hybrid system model-checking tools for dealing with hybrid planning domains. As a case study, we use the SpaceEx model checker, showing how we can address PDDL+ domains that are out of the scope of state-of-the-art planners.

While MDPs are powerful tools for modeling sequential decision making problems under uncertainty, they are sensitive to the accuracy of their parameters. MDPs with uncertainty in their parameters are called Uncertain MDPs. In this paper, we introduce a general framework that allows off-the-shelf MDP algorithms to solve Uncertain MDPs by planning based on currently available information and replan if and when the problem changes. We demonstrate the generality of this approach by showing that it can use the VI, TVI, ILAO*, LRTDP, and UCT algorithms to solve Uncertain MDPs. We experimentally show that our approach is typically faster than replanning from scratch and we also provide a way to estimate the amount of speedup based on the amount of information being reused.

We present work toward computationally defining a model of narrative comprehension vis-à-vis memory of narrative events, via an automated planning knowledge representation, capable of being used in a narrative generation context.