A game-theoretic approach to timeline-based planning with uncertainty
This work addresses the problem of handling complex nondeterminism in timeline-based planning for AI systems, representing an incremental advance by generalizing existing methods.
The paper tackles the limitation of flexible plans in timeline-based planning, which only handle temporal uncertainty, by proposing a game-theoretic approach that uniformly handles both temporal uncertainty and nondeterminism, showing that winning strategies are more general than control strategies and the existence problem is decidable with doubly exponential space complexity.
In timeline-based planning, domains are described as sets of independent, but interacting, components, whose behaviour over time (the set of timelines) is governed by a set of temporal constraints. A distinguishing feature of timeline-based planning systems is the ability to integrate planning with execution by synthesising control strategies for flexible plans. However, flexible plans can only represent temporal uncertainty, while more complex forms of nondeterminism are needed to deal with a wider range of realistic problems. In this paper, we propose a novel game-theoretic approach to timeline-based planning problems, generalising the state of the art while uniformly handling temporal uncertainty and nondeterminism. We define a general concept of timeline-based game and we show that the notion of winning strategy for these games is strictly more general than that of control strategy for dynamically controllable flexible plans. Moreover, we show that the problem of establishing the existence of such winning strategies is decidable using a doubly exponential amount of space.