Systems
Agents
- IndiGolog: A situation-calculus based high-level programming language for autonomous agents that sense their environment and do planning as they operate. Instead of classical planning, it supports high-level program execution. Part of the University of Toronto Golog family of high-level programming and planning languages.
Automated Planning
FOND planning
- FOND+-ASP: A FOND planner in Answer Set Programing handling conditional fairness. Best Paper Award at ICAPS'21.
- CFOND-ASP: A FOND planner in Answer Set Programming to extract compact controllers (inspired by FOND-SAT and PRP).
- fond2allout: All-outcome lifted determinizer for FOND domains in PDDL.
- This is different from the determinizers built-in in planners like PRP, because they are modifications of FD translate script to encode PDDL problems into SAS representation. The
fond2alloutsystem can determinize a domain without a problem and still produces a PDDL.
- This is different from the determinizers built-in in planners like PRP, because they are modifications of FD translate script to encode PDDL problems into SAS representation. The
Agent Planning Programs
- PP-TRAPPER: Agent Planning Programs via Trap Reasoning. Code for the IJCAI'17 paper.
Path Planning (used for GR too)
- p4: p4 (aka the “Python Path Planning Project is a Python-based path planning framework and simulator, useful to prototype, evaluate, and benchmark path planning algorithms. It was also used for Goal Wecognition work.
- Apparate Simulator: Java-based tool to code, simulate, and visualise various path planning algorithms, such as A*, Real-time A* or incremental path planners like D* or Moving Target. Importantly, APPARATE is a suitable tool for testing incremental algorithms; it allows dynamic changes to the environment (like adding water areas or moving the target location), via a script or the GUI interface.
Plan Flexibility
- MRR: Implementation of a MaxSAT-based technique for finding minimum deorderings and reorderings of a partial order plan. It can be configured to either find a minimum deorder or minimum reorder as per Muise et al., or a minimum reinstantiated deorder or minimum reiinstantiated reorder as per Waters et al.
- MKTR: MKTR is an application for finding re-instantiations of a plan, i.e., alternative plans which achieve the same goal and use the same actions, but differ in the order of those actions and how the variables in those actions have been bound.
Goal Recognition
xxx
Others
- tapp-fixture: Python scripts to scrape sport competition games from PlayHQ, and produce CSV files for importing team schedules into Stack TeamApps, a popular sport app used in Australia. Script is being used by domestic and rep basketball teams in Victoria.
- pynbl: Python scripts to extract statistics from Australian NBL league.