Space-Time Planner
Skip down to diagrams

Overview

The motivating task for the Space-Time Planner is scheduling of talks at a conference, and the related allocation of rooms and equipment, in a crisis situation. The scenario contains an initial schedule and then an unexpected major change in space availability; for example, a building closes. There is also a continuous stream of minor changes; for example, schedule changes and unforeseen equipment needs.

The learning-based system begins with information elicitation and ends with the learning of recombinant plans and strategies (part of future research plans).

Technical Details

The system currently identifies critical missing knowledge, sends related questions to users, and improves the world model based on users' answers. It also analyzes old elicitation logs and creates rules for "static" generation of useful questions. These rules enable the system to ask critical questions before scheduling. Finally, the system analyzes known requirements and user preferences, and creates rules for generating default preferences. These rules enable the system to make reasonable assumptions about unknown requirements and preferences.

Future work (see second screen shot) will focus on:

- Learning of control rules for high-level planning and elicitation strategies

- Recombinant actions sequences and plans

- Automated selection reasoning and learning strategies from a library

- Advanced mechanism for dealing with uncertainty and hypothetical scenarios

Diagrams



back to top