Miquel Ramirez

Goal recognition over POMDPs: inferring the intention of a POMDP agent

Plan recognition is the problem of inferring the goals and plans of an agent from partial observations of her behavior. Recently, it has been shown that the problem can be formulated and solved using planners, reducing plan recognition to plan generation. In this work, we extend this model-based approach to plan recognition to the POMDP setting, where actions are stochastic and states are partially observable. The task is to infer a probability distribution over the possible goals of an agent whose behavior results from a POMDP model. The POMDP model is shared between agent and observer except for the true goal of the agent that is hidden to the observer. The observations are action sequences O that may contain gaps as some or even most of the actions done by the agent may not be observed. We show that the posterior goal distribution P(G|O) can be computed from the value function VG(b) over beliefs b generated by the POMDP planner for each possible goal G. Some extensions of the basic framework are discussed, and a number of experiments are reported.

Structural relaxations by variable renaming and their compilation for solving MinCostSAT

Searching for optimal solutions to a problem using lower bounds obtained from a relaxation is a common idea in Heuristic Search and Planning. In SAT and CSPs, however, explicit relaxations are seldom used. In this work, we consider the use of explicit relaxations for solving MinCostSAT, the problem of finding a minimum cost satisfying assignment. We start with the observation that while a number of SAT and CSP tasks have a complexity that is exponential in the treewidth, such models can be relaxed into weaker models of bounded treewidth by a simple form of variable renaming. The relaxed models can then be compiled in polynomial time and space, so that their solutions can be used effectively for pruning the search in the original problem. We have implemented a MinCostSAT solver using this idea on top of two off-the-shelf tools, a d-DNNF compiler that deals with the relaxation, and a SAT solver that deals with the search. The results over the entire suite of 559 problems from the 2006 Weighted Max-SAT Competition are encouraging: SR(w), the new solver, solves 56% of the problems when the bound on the relaxation treewidth is set to w = 8, while Toolbar, the winner, solves 73% of the problems, Lazy, the runner up, 55%, and MinCostChaff, a recent MinCostSAT solver, 26%. The relation between the proposed relaxation method and existing structural solution methods such as cutset decomposition and derivatives such as mini-buckets is also discussed.

Real--Time UAV Maneuvering via Automated Planning in Simulations

The automatic generation of realistic behaviour such as tactical intercepts for Unmanned Aerial Vehicles (UAV) in air combat is a challenging problem. State-of-the-art solutions propose hand–crafted algorithms and heuristics whose performance depends heavily on the initial conditions and specific aerodynamic characteristics of the UAVs involved. This demo shows the ability of domain–independent planners, embedded into simulators, to generate on–line, feed–forward, control signals that steer simulated aircraft as best suits the situation

Extending Classical Planning with State Constraints: Heuristics and Search for Optimal Planning

We present a principled way of extending a classical AI planning formalism with systems of state constraints, which relate – sometimes determine – the values of variables in each state traversed by the plan. This extension occupies an attractive middle ground between expressivity and complexity. It enables modelling a new range of problems, as well as formulating more efficient models of classical planning problems. An example of the former is planning-based control of networked physical systems – power networks, for example – in which a local, discrete control action can have global effects on continuous quantities, such as altering flows across the entire network. At the same time, our extension remains decidable as long as the satisfiability of sets of state constraints is decidable, including in the presence of numeric state variables, and we demonstrate that effective techniques for costoptimal planning known in the classical setting – in particular, relaxation-based admissible heuristics – can be adapted to the extended formalism. In this paper, we apply our approach to constraints in the form of linear or non-linear equations over numeric state variables, but the approach is independent of the type of state constraints, as long as there exists a procedure that decides their consistency. The planner and the constraint solver interact through a well-defined, narrow interface, in which the solver requires no specialisation to the planning context. Furthermore, we present an admissible search algorithm – a variant of A – that is able to make use of additional information provided by the search heuristic, in the form of preferred actions. Although preferred actions have been widely used in satisficing planning, we are not aware of any previous use of them in optimal planning. c ©2018 AI Access Foundation. All rights reserved. Haslum, Ivankovic, Raḿırez, Gordon, Thiébaux, Shivashankar & Nau