In many practical domains, planning systems are required to reason about durative actions. A common assumption in the literature is that the executor is allowed to decide the duration of each action. However, this assumption may be too restrictive for applications. In this paper, we tackle the problem of temporal planning with uncontrollable action durations. We show how to generate robust plans,that guarantee goal achievement despite the uncontrollability of the actual duration of the actions. We extend the state-space temporalplanning framework, integrating recent techniques for solving temporalproblems under uncertainty. We discuss different ways of lifting the total order plans generated by the heuristic search to partial orderplans, showing (in)completeness results for each of them. We implemented our approach on top of COLIN, a state-of-the-art planner. An experimental evaluation over several benchmark problems shows the practical feasibility of the proposed approach.

Fully observable decision-theoretic planning problems are commonly modeled as stochastic shortest path (SSP) problems. For this class of planning problems, heuristic search algorithms (including LAO*, RTDP, and related algorithms), as well as the value iteration algorithm on which they are based, lack an efficient test for convergence to an ε-optimal policy (except in the special case of discounting). We introduce a simple and efficient test for convergence that applies to SSP problems with positive action costs. The test can detect whether a policy is proper, that is, whether it achieves the goal state with probability 1. If proper, it gives error bounds that can be used to detect convergence to an ε-optimal solution. The convergence test incurs no extra overhead besides computing the Bellman residual, and the performance guarantee it provides substantially improves the utility of this class of planning algorithms.

Many AI applications involve the interaction of multiple autonomous agents, requiring those agents to reason about their own beliefs, as well as those of other agents. However, planning involving nested beliefs is known to be computationally challenging. In this work, we address the task of synthesizing plans that necessitate reasoning about the beliefs of other agents. We plan from the perspective of a single agent with the potential for goals and actions that involve nested beliefs, non-homogeneous agents, co-present observations, and the ability for one agent to reason as if it were another. We formally characterize our notion of planning with nested belief, and subsequently demonstrate how to automatically convert such problems into problems that appeal to classical planning technology. Our approach represents an important first step towards applying the well-established field of automated planning to the challenging task of planning involving nested beliefs of multiple agents.

We propose revisions to the research agenda in Automated Planning. The proposal is based on a review of the role of the Planning Domain Definition Language (PDDL) in the activities of the AI planning community and the impact of PDDL on parts of its research agenda. We specifically show how specific properties of PDDL have impacted research on planning, by putting emphasis on certain research topics and complicating others. We argue that the development of more advanced modeling languages would be — analogously to the impact PDDL has had — a low overhead and smooth route for the ICAPS community shift its research focus to increasingly promising and relevant research topics.

Merge-and-shrink heuristics crucially rely on effective reduction techniques, such as bisimulation-based shrinking, to avoid the combinatorial explosion of abstractions. We propose the concept of factored symmetries for merge-and-shrink abstractions based on the established concept of symmetry reduction for state-space search. We investigate under which conditions factored symmetry reduction yields perfect heuristics and discuss the relationship to bisimulation. We also devise practical merging strategies based on this concept and experimentally validate their utility.

For my dissertation I am focusing on non-classical planning for robotic applications. Much classical planning research relies on assumptions that do not hold in real world robotics applications. In many cases the entire world state is not known in advance and the events that occur in the future can not be known with certainty. Robots operating in the real world also need to be responsive and react to dynamic obstacles and events.

We propose an auction algorithm to allocate tasks that have temporal constraints to cooperative robots. Temporal constraints are expressed as time windows, within which a task must be executed. There are no restrictions on the time windows, which are allowed to overlap. Robots model their temporal constraints using a simple temporal network, enabling them to maintain consistent schedules. When bidding on a task, a robot takes into account its own current commitments and an optimization objective, which is to minimize the time of completion of the last task alone or in combination with minimizing the distance traveled. The algorithm works both when all the tasks are known upfront and when tasks arrive dynamically. We show the performance of the algorithm in simulation with different numbers of tasks and robots, and compare it with a baseline greedy algorithm and a state-of-the-art auction algorithm. Our algorithm is computationally frugal and consistently allocates more tasks than the competing algorithms.

Scheduling is integral to many real-world logistics problems. It can be as simple as catching the bus in the morning, or as complex as assembling a commercial airliner. While simple applications render scheduling tools trivial, these tools have not been widely adopted for complex scenarios either. The larger the scenario, the greater the temporal uncertainty throughout the system, and many schedulers do not consider the probabilistic uncertainty in actions' durations. Figure 1: The role of scheduling in a plannning and execution This makes them brittle to temporal disturbances or architecture. In this architecture, the planner and scheduler first generate Figure 1 diagrams the layers of reasoning for a planning executive a plan and scheduling policy offline, which the dispatcher to map logistical goals into real-world actions.

In a recent position paper in Artificial Intelligence, we argued that the automated planning research literature has underestimated the importance and difficulty of deliberative acting, which is more than just interleaving planning and execution. We called for more research on the AI problems that emerge when attempting to integrate acting with planning. To provide a basis for such research, it will be important to have a formalization of acting that can be useful in practice. This is needed in the same way that a formal account of planning was necessary for research on planning. We describe some first steps toward developing such a formalization, and invite readers to carry out research along this line.

In this paper we present a plan-plan distance metric based on Kolmogorov(Algorithmic) complexity. Generating diverse sets of plans is useful for task ssuch as probing user preferences and reasoning about vulnerability to cyberattacks. Generating diverse plans, and comparing different diverse planning approaches requires a domain-independent, theoretically motivated definition of the diversity distance between plans. Previously proposed diversity measures are not theoretically motivated, and can provide inconsistent results on the sameplans. We define the diversity of plans in terms of how surprising one plan is givenanother or, its inverse, the conditional information in one plan givenanother. Kolmogorov complexity provides a domain independent theory of conditional information. While Kolmogorov complexity is not computable, a related metric, Normalized Compression Distance (NCD), provides a well-behaved approximation. In this paper we introduce NCD as an alternative diversity metric, and analyze its performance empirically, in comparison with previous diversity measures, showing strengths and weaknesses of each.We also examine the use of different compressor sin NCD. We show how NCD can be used to select a training set for HTN learning,giving an example of the utility of diversity metrics. We conclude withsuggestions for future work on improving, extending, and applying it to serve new applications.