Thomas Allen

AAAI Conference 2016 Conference Paper

Generating CP-Nets Uniformly at Random

• Thomas Allen
• Judy Goldsmith
• Hayden Justice
• Nicholas Mattei
• Kayla Raines

Conditional preference networks (CP-nets) are a commonly studied compact formalism for modeling preferences. To study the properties of CP-nets or the performance of CP-net algorithms on average, one needs to generate CP-nets in an equiprobable manner. We discuss common problems with naı̈ve generation, including sampling bias, which invalidates the base assumptions of many statistical tests and can undermine the results of an experimental study. We provide a novel algorithm for provably generating acyclic CP-nets uniformly at random. Our method is computationally efficient and allows for multi-valued domains and arbitrary bounds on the indegree in the dependency graph.

AAAI Conference 2016 Conference Paper

Model AI Assignments 2016

• Todd Neller
• Laura Brown
• James Marshall
• Lisa Torrey
• Nate Derbinsky
• Andrew Ward
• Thomas Allen
• Judy Goldsmith

AAAI Conference 2014 Conference Paper

Making CP-Nets (More) Useful

• Thomas Allen

ICRA Conference 2011 Conference Paper

The Time-Optimal Planning and Execution problem

• Thomas Allen
• Steve Scheding

This paper introduces the Time-Optimal Planning and Execution (TOPE) problem, in which the aim is to minimise the total planning and execution time required to achieve a goal. The TOPE process is derived and shown to be capable of solving this problem in dynamic state spaces, by continuously calculating the optimum value of any system parameters that can affect this total time. Procedures are presented to apply this process to an existing replanning system, and to determine its required accuracy and timeliness. It is shown that the TOPE process can yield lower total times than other planning systems if these requirements are met.

ICRA Conference 2009 Conference Paper

Dynamic path planning with multi-agent data fusion - The Parallel Hierarchical Replanner

• Thomas Allen
• Andrew John Hill
• James Patrick Underwood
• Steve Scheding

The design of a hierarchical planning system in which each level operates in parallel and communicates asynchronously is presented. It is shown that this Parallel Hierarchical Replanner is both reactive, and as close to optimal over all information in the state space as is possible given finite computational power. A comparison with three other hierarchical methods is presented, which demonstrates that for scenarios in which the time taken to achieve a mission goal is of greater importance than the cost incurred, this approach has better performance than related methods in the literature.