Planning is vital to a wide range of domains, including robotics, military strategy, logistics, itinerary generation and more, that both humans and computers find difficult. Collaborative planning holds the promise of greatly improving performance on these tasks by leveraging the strengths of both humans and automated planners. However, this requires formalizing the problem domain and input, which must be done by hand, a priori, restricting its use in general real-world domains. We propose using a real-time crowd of workers to simultaneously solve the planning problem, formalize the domain, and train an automated system. As plans are developed, the system is able to learn the domain, and contribute larger segments of work.

Diverse planning consists of generating multiple different solutions for the same planning problem. I explore solution diversity, based on quantitative (domain-independent) and qualitative (domain-dependent) distance metrics, in deterministic and nondeterministic planning domains.

UCT, the premier method for solving games such as Go, is also becoming the dominant algorithm for probabilistic planning. Out of the five solvers at the International Probabilistic Planning Competition (IPPC) 2011, four were based on the UCT algorithm. However, while a UCT-based planner, PROST, won the contest, an LRTDP-based system, Glutton, came in a close second, outperforming other systems derived from UCT. These results raise a question: what are the strengths and weaknesses of LRTDP and UCT in practice? This paper starts answering this question by contrasting the two approaches in the context of finite-horizon MDPs. We demonstrate that in such scenarios, UCT's lack of a sound termination condition is a serious practical disadvantage. In order to handle an MDP with a large finite horizon under a time constraint, UCT forces an expert to guess a non-myopic lookahead value for which it should be able to converge on the encountered states. Mistakes in setting this parameter can greatly hurt UCT's performance. In contrast, LRTDP's convergence criterion allows for an iterative deepening strategy. Using this strategy, LRTDP automatically finds the largest lookahead value feasible under the given time constraint. As a result, LRTDP has better performance and stronger theoretical properties. We present an online version of Glutton, named Gourmand, that illustrates this analysis and outperforms PROST on the set of IPPC-2011 problems.

One of the natural approaches for selecting actions in very From this perspective, an algorithm like RTDP fails on two large state spaces is by performing a limited amount of grounds: first, RTDP does not appear to make best use of lookahead. In the contexts of discounted MDPs, Kearns, short time windows in large state spaces; second, and more Mansour, and Ng have shown that near to optimal actions importantly, RTDP can use admissible heuristics but not informed can be selected by considering a sampled lookahead tree that base policies. On the other hand, algorithms like Policy is sufficiently sparse, whose size depends on the discount Iteration (Howard 1971), deliver all of these features except factor and the suboptimality bound but not on the number of one: they are exhaustive, and thus even to get started, problem states (Kearns, Mansour, and Ng 1999). The UCT they need vectors with the size of the state space. At the algorithm (Kocsis and Szepesvári 2006) is a version of this same time, while there are non-exhaustive versions of (asynchronous) form of Monte Carlo planning, where the lookahead trees Value Iteration such as RTDP, there are no similar are not grown depth-first but'best-first', following a selection'focused' versions of Policy Iteration ensuring anytime optimality.

We address the problem of generalizing a plan to maximize To date we have focused on generalizing partial-order plans the flexibility and robustness with which an agent can execute (POPs). We described how to generalize a POP for online it. A key area of automated planning is the study of how execution, and laid the groundwork for the approach our research to generate a plan for an agent to execute. The plan itself will follow (Muise, McIlraith, and Beck 2011a). We may take on many forms: a sequence of actions, a partial ordering also presented a method to optimally relax a sequential plan over a set of actions, or a procedure-like description to a POP (Muise, Mcilraith, and Beck 2011b), providing a of what the agent should do.

In hierarchical planning, landmarks are tasks that occur on any search path leading from the initial plan to a solution. In this work, we present novel domain-independent planning strategies based on such hierarchical landmarks. Our empirical evaluation on four benchmark domains shows that these landmark-aware strategies outperform established search strategies in many cases.

Tractability analysis in terms of the causal graphs of planning problems has emerged as an important area of research in recent years, leading to new methods for the derivation of domain-independent heuristics (Katz and Domshlak 2010). Here we continue this work, extending our knowledge of the frontier between tractable and NP-complete fragments. We close some gaps left in previous work, and introduce novel causal graph fragments that we call the hourglass and semifork, for which under certain additional assumptions optimal planning is in P. We show that relaxing any one of the restrictions required for this tractability leads to NP-complete problems. Our results are of both theoretical and practical interest, as these fragments can be used in existing frameworks to derive new abstraction heuristics. Before they can be used, however, a number of practical issues must be addressed. We discuss these issues and propose some solutions.

The early classifications of the computational complexity of planning under various restrictions in STRIPS (Bylander) and SAS+ (Bäckström and Nebel) have influenced following research in planning in many ways. We go back and reanalyse their subclasses, but this time using the more modern tool of parameterized complexity analysis. This provides new results that together with the old results give a more detailed picture of the complexity landscape. We demonstrate separation results not possible with standard complexity theory, which contributes to explaining why certain cases of planning have seemed simpler in practice than theory has predicted. In particular, we show that certain restrictions of practical interest are tractable in the parameterized sense of the term, and that a simple heuristic is sufficient to make a well-known partial-order planner exploit this fact.

We describe a new sound and complete method forcompiling contingent planning problems with sensingactions into classical planning. Our method encodesconditional plans within a linear, classicalplan. This allows our planner, MPSR, to reasonabout multiple future outcomes of sensing actions,and makes it less susceptible to dead-ends. MPRS,however, generates very large classical planningproblems. To overcome this, we use an incompletevariant of the method, based on state sampling,within an online replanner. On most currentdomains, MPSR finds plans faster, although itsplans are often longer. But on a new challengingvariant of Wumpus with dead-ends, it finds smallerplans, faster, and scales much better.

When an agent plans a sequence of actions, some unexpected events may occur during the execution of these actions. These unexpected events may prevent the agent to replan and achieve its goal. In this work, our purpose is to recover from plan execution failures by reasoning the causes of these faulties. We combine the TLPlan forward chaining temporal planner with the PROBCOG reasoning tool in order to handle failures. It is also quite important to decide whether the failure we are dealing with is permanent. We propose that inferring some properties of the failure source helps us handle failures and determine the failure types.