Engineering complete planning domain descriptions is often very costly because of human error or lack of domain knowl- edge. Learning complete domain descriptions is also very challenging because many features are irrelevant to achieving the goals and data may be scarce. We present a planner and agent that respectively plan and act in incomplete domains by i) synthesizing plans to avoid execution failure due to ignorance of the domain model, and ii) passively learning about the domain model during execution to improve later re-planning attempts. Our planner DeFault is the first to reason about a domain’s incompleteness to avoid potential plan failure. DeFault computes failure explanations for each action and state in the plan and counts the number of interpretations of the incomplete domain where failure will occur. We show that DeFault performs best by counting prime implicants (failure diagnoses) rather than propositional models. Our agent Goalie learns about the preconditions and effects of incompletely-specified actions while monitoring its state and, in conjunction with DeFault plan failure explanations, can diagnose past and future action failures. We show that by reasoning about incompleteness (as opposed to ignoring it) Goalie fails and re-plans less and executes fewer actions.

Consider the task of a mobile robot autonomously navigating through an environment while detecting and mapping objects of interest using a noisy object detector. The robot must reach its destination in a timely manner, but is rewarded for correctly detecting recognizable objects to be added to the map, and penalized for false alarms. However, detector performance typically varies with vantage point, so the robot benefits from planning trajectories which maximize the efficacy of the recognition system. This work describes an online, any-time planning framework enabling the active exploration of possible detections provided by an off-the-shelf object detector. We present a probabilistic approach where vantage points are identified which provide a more informative view of a potential object. The agent then weighs the benefit of increasing its confidence against the cost of taking a detour to reach each identified vantage point. The system is demonstrated to significantly improve detection and trajectory length in both simulated and real robot experiments.

In this paper optimal state space planning is parallelized by exploiting the processing power of a graphics card. The two exploration steps, namely selecting the actions to be applied and generating the successors, are performed on a graphics processing unit. Duplicate detection, however, is delayed to be executed on the central processing unit. Multiple cores are employed to bypass main memory latency. To increase processing speed for exact duplicate detection, the hash tables are lock-free. Moreover, a bucket-based representation enhances the concurrent distribution of frontier states. The planner supports cost-first exploration and is able to deal with a considerable fraction of current PDDL, including numerical state variables, complex objective functions, and goal preferences. It can maximize the net-benefit. Experimental findings show visible performance gains especially for larger benchmark problems.

We consider the problem of finding generalized plans for situations where the number of objects may be unknown and unbounded during planning. The input is a domain specification, a goal condition, and a class of concrete problem instances or initial states to be solved, expressed in an abstract first-order representation. Starting with an empty generalized plan, our overall approach is to incrementally increase the applicability of the plan by identifying a problem instance that it cannot solve, invoking a classical planner to solve that problem, generalizing the obtained solution and merging it back into the generalized plan. The main contributions of this paper are methods for (a) generating and solving small problem instances not yet covered by an existing generalized plan, (b) translating between concrete classical plans and abstract plan representations, and (c) extending partial generalized plans and increasing their applicability. We analyze the theoretical properties of these methods, prove their correctness, and illustrate experimentally their scalability. The resulting hybrid approach shows that solving only a few, small, classical planning problems can be sufficient to produce a generalized plan that applies to infinitely many problems with unknown numbers of objects.

This paper presents a generalization of the coupled-task scheduling problem introduced by Shapiro, where considered tasks are subject to incompatibility constraint depicted by an undirected graph. The motivation of this problem comes from data acquisition and processing in a mono-processor torpedo used for underwater exploration. As we add the compatibility graph, we focus on complexity of the problem, and more precisely on the border between P and NP-completeness when some other input parameters are restricted (e.g. the ratio between the durations of the two sub-tasks composing a task): we adapt the global visualization of the complexity of scheduling problems with coupled-task given by Orman and Potts to our problem, determine new complexity results, and thus propose a new visualization including incompatibility constraint. In the end, we give a new polynomial-time approximation algorithm result which completes previous works.

We present the first effective SAT heuristics for planning with expressive planning languages such as ADL. Recently, SAT heuristics for STRIPS planning have been introduced. In this work we show that the basic ideas in the heuristic can be generalized to actions with conditional effects but without disjunction, and that disjunction requires a more fundamental analysis of the STRIPS heuristic, which, despite complications, will still lead to a natural heuristic which can be implemented efficiently. The experimental analysis shows substantial and systematic improvements over the state of the art in planning with SAT with ADL.

We present a software tool that is able to automatically translate an NP problem into a STRIPS problem such that the former problem has a solution iff the latter has one, a solution for the latter can be transformed into a solution for the former, and all this can be done efficiently. Moreover, the tool is built such that it only produces problems that belong to a fragment of STRIPS that is solvable in non-deterministic polynomial time, a fact that guarantees that the whole approach is not an overkill. This tool has interesting applications. For example, with the advancement of planning technology, it can be used as an off-the-shelf method to solve general NP problems with the help of planners and to automatically generate benchmark problems of known complexity in a systematic and controlled manner. Another interesting contribution is related to the area of Knowledge Engineering in which one of the goals is to devise automatic methods for using the available planning technology to solve real-life problems.

Considering directed graphs on n vertices and m edges with real (possibly negative) weights, we present two new, efficient algorithms for computing all-pairs shortest paths (APSP). These algorithms make use of directed path consistency (DPC) along a vertex ordering d. The algorithms run in O(n 2 w d ) time, where w d is the graph width induced by this vertex ordering. For graphs of constant treewidth, this yields O(n 2 ) time, which is optimal. On chordal graphs, the algorithms run in O(nm) time. We show empirically that also in many general cases, both constructed and from realistic benchmarks, the algorithms often outperform Johnson's algorithm, which represents the current state of the art with a run time of O(nm + n 2 log n). These algorithms can be used for temporal and spatial reasoning, e.g. for the Simple Temporal Problem (STP), which underlines its relevance to the planning and scheduling community.

We consider an application of scheduling to hardware-accelerated functional verification, a massively-parallel computational paradigm used in the simulation of complex integrated circuits. Our domain requires the compilation of logical primitives into a set of instruction memories that optimize the concurrency and communication between tightly synchronized processing units. The scheduling process is burdened by a complex model in which all logical dependencies must be resolved by a dynamic network of routes that compete for sparsely distributed resources. We describe a series of optimization steps that cooperate to minimize simulation depth while scaling to problem sizes on the order of a billion gates. Our approach targets an industrial acceleration architecture containing 262,144 parallel processors.

We define a probe to be a single action sequence computedgreedily from a given state that either terminates in the goalor fails. We show that by designing these probes carefullyusing a number of existing and new polynomial techniquessuch as helpful actions, landmarks, commitments, and con-sistent subgoals, a single probe from the initial state solvesby itself 683 out of 980 problems from previous IPCs, a num-ber that compares well with the 627 problems solved by FFin EHC mode, with similar times and plan lengths. We alsoshow that by launching one probe from each expanded statein a standard greedy best first search informed by the addi-tive heuristic, the number of problems solved jumps to 900(92%), as opposed to FF that solves 827 problems (84%),and LAMA that solves 879 (89%). The success of probessuggests that many domains can be solved easily once a suit-able serialization of the landmarks is found, an observationthat may open new connections between recent work in plan-ning and more classical work concerning goal serializationand problem decomposition in planning and search.