Monte-Carlo Tree Search (MCTS) algorithms estimate the value of MDP states based on rewards received by performing multiple random simulations. MCTS algorithms can use different strategies to aggregate these rewards and provide an estimation for the states’ values. The most common aggregation method is to store the mean reward of all simulations. Another common approach stores the best observed reward from each state. Both of these methods have complementary benefits and drawbacks. In this paper, we show that both of these methods are biased estimators for the real expected value of MDP states. We propose an hybrid approach that uses the best reward for states with low noise, and otherwise uses the mean. Experimental results on the Sailing MDP domain show that our method has a considerable advantage when the rewards are drawn from a noisy distribution.

The Grid-Based Path Planning Competition has just completed its third iteration. The entriesused in the competition have improved significantly during this time, changing the view ofthe state of the art of grid-based pathfinding. Furthermore, the entries from the competition have beenmade publicly available, improving the ability of researchers to compare their work. Thispaper summarizes the entries to the 2014 competition, presents the 2014 competition results,and talks about what has been learned and where there is room for improvement.

Fork decoupling is a recent approach to exploiting problem structure in state space search. The problem is assumed to take the form of a fork, where a single (large) center component provides preconditions for several (small) leaf components. The leaves are then conditionally independent in the sense that, given a fixed center path p, the compliant leaf moves - those leaf moves enabled by the preconditions supplied along p - can be scheduled independently for each leaf. Fork-decoupled state space search exploits this through conducting a regular search over center paths, augmented with maintenance of the compliant paths for each leaf individually. We herein show that the same ideas apply to much more general star-topology structures, where leaves may supply preconditions for the center, and actions may affect several leaves simultaneously as long as they also affect the center. Our empirical evaluation in planning, super-imposing star topologies by automatically grouping the state variables into suitable components, shows the merits of the approach.

We compare five any-angle path-planning algorithms, Theta*, Block A*, Field D*, ANYA, and Any-Angle Subgoal Graphs in terms of solution quality and runtime. Any-angle path-planning is a fairly new research area, and no direct comparison exists between these algorithms. We implement each algorithm from scratch and use similar implementations to provide a fair comparison.

Definition of an accurate system model for Automated Planner (AP) is often impractical, especially for real-world problems. Conversely, off-the-shelf planners fail to scale up and are domain dependent. These drawbacks are inherited from conventional transition systems such as Finite State Machines (FSMs) that describes the action-plan execution generated by the AP. On the other hand, Behavior Trees (BTs) represent a valid alternative to FSMs presenting many advantages in terms of modularity, reactiveness, scalability and domain-independence. In this paper, we propose a model-free AP framework using Genetic Programming (GP) to derive an optimal BT for an autonomous agent to achieve a given goal in unknown (but fully observable) environments. We illustrate the proposed framework using experiments conducted with an open source benchmark Mario AI for automated generation of BTs that can play the game character Mario to complete a certain level at various levels of difficulty to include enemies and obstacles.

We aim to reduce the burden of programming and deploying autonomous systems to work in concert with people in time-critical domains such as military field operations and disaster response. Deployment plans for these operations are frequently negotiated on-the-fly by teams of human planners. A human operator then translates the agreed-upon plan into machine instructions for the robots. We present an algorithm that reduces this translation burden by inferring the final plan from a processed form of the human team's planning conversation. Our hybrid approach combines probabilistic generative modeling with logical plan validation used to compute a highly structured prior over possible plans, enabling us to overcome the challenge of performing inference over a large solution space with only a small amount of noisy data from the team planning session. We validate the algorithm through human subject experimentations and show that it is able to infer a human team's final plan with 86% accuracy on average. We also describe a robot demonstration in which two people plan and execute a first-response collaborative task with a PR2 robot. To the best of our knowledge, this is the first work to integrate a logical planning technique within a generative model to perform plan inference.

Dr. Ford is being recognized Outstanding Program

This paper describes a challenging, real-world planning problem within the context of a NASA mission called LADEE (Lunar Atmospheric Dust Environment Explorer). We present the approach taken to reduce the complexity of the activity planning task in order to effectively perform it within the time pressures imposed by the mission requirements. One key aspect of this approach is the design of the activity planning process based on principles of problem decomposition and planning abstraction levels. The second key aspect is the mixed-initiative system developed for this task, called LASS (LADEE Activity Scheduling System). The primary challenge for LASS was representing and managing the science constraints that were tied to key points in the spacecraft’s orbit, given their dynamic nature due to the continually updated orbit determination solution.

While model-based reinforcement learning is often studied under the assumption that a fully accurate model is contained within the model class, this is rarely true in practice. When the model class may be fundamentally limited, it can be difficult to obtain theoretical guarantees. Under some conditions the DAgger algorithm promises a policy nearly as good as the plan obtained from the most accurate model in the class, but only if the planning algorithm is near-optimal, which is also rarely the case in complex problems. This paper explores the interaction between DAgger and Monte Carlo planning, specifically showing that DAgger may perform poorly when coupled with a sub-optimal planner. A novel variation of DAgger specifically for use with Monte Carlo planning is derived and is shown to behave far better in some cases where DAgger fails.

When scheduling tasks for field-deployable systems, our solutions must be robust to the uncertainty inherent in the real world. Although human intuition is trusted to balance reward and risk, humans perform poorly in risk assessment at the scale and complexity of real world problems. In this paper, we present a decision aid system that helps human operators diagnose the source of risk and manage uncertainty in temporal problems. The core of the system is a conflict-directed relaxation algorithm, called Conflict-Directed Chance-constraint Relaxation (CDCR), which specializes in resolving over-constrained temporal problems with probabilistic durations and a chance constraint bounding the risk of failure. Given a temporal problem with uncertain duration, CDCR proposes execution strategies that operate at acceptable risk levels and pinpoints the source of risk. If no such strategy can be found that meets the chance constraint, it can help humans to repair the over-constrained problem by trading off between desirability of solution and acceptable risk levels. The decision aid has been incorporated in a mission advisory system for assisting oceanographers to schedule activities in deep-sea expeditions, and demonstrated its effectiveness in scenarios with realistic uncertainty.