SIFT researcher wins Best Student Paper Award at ICAPS
SIFT researcher J. Benton won the Best Student Paper Award at ICAPS 2012 called "Temporal Planning with Preferences and Time-dependent Costs," also authored by Amanda Coles and Andrew Coles. This paper is about handling temporal planning problems where the quality of a plan has to do with when certain goals are achieved. The paper's abstract is after the break.
J. was also second author on a short paper presented at ICAPS entitled "Anticipatory On-line Planning," also authored by Ethan Burns, Wheeler Ruml, Minh Do, and Sungwook Yoon.
Abstract for "Temporal Planning with Preferences and Time-dependent Costs"
Temporal planning methods usually focus on the objective of minimizing makespan. Unfortunately, this misses a large class of planning problems where it is important to consider a wider variety of temporal and non-temporal preferences, making makespan a lower-order concern. In this paper we consider modeling and reasoning with plan quality metrics that are not directly correlated with plan makespan, building on the planner POPF. We begin with the preferences defined in PDDL3, and present a mixed integer programming encod- ing to manage the the interaction between the hard tempo- ral constraints for plan steps, and soft temporal constraints for preferences. To widen the support of metrics that can be expressed directly in PDDL, we then discuss an extension to soft-deadlines with continuous cost functions, avoiding the need to approximate these with several PDDL3 discrete-cost preferences. We demonstrate the success of our new planner on the benchmark temporal planning problems with prefer- ences, showing that it is the state-of-the-art for such prob- lems. We then analyze the benefits of reasoning with contin- uous (versus discretized) models of domains with continuous cost functions, showing the improvement in solution qual- ity afforded through making the continuous cost function di- rectly available to the planner.