Skip to main content

Least Commitment Flexible Partial-Order Planning

Primary supervisor

Buser Say

Co-supervisors


Automated planning is the reasoning side of acting in Artificial Intelligence. An autonomous agent must plan, i.e., select and order its actions, to achieve its goals as best as possible. A partial-order plan compactly represents a set of plans which provides the agent some execution flexibility. This project will focus on building a decomposition-based least commitment flexible partial-order planner.

Student cohort

Double Semester

URLs/references

Mathematical modeling based approaches to:

  • automated planning heuristics [1,2,3],
  • partial-order relaxation [4,5,6],
  • decomposition based automated planning [7].

constitute a good starting point for literature review.

[1] Tom Bylander. A Linear Programming Heuristic for Optimal Planning. 1997.

[2] Menkes van den Briel, J. Benton, Subbarao Kambhampati, Thomas Vossen. An LP-Based Heuristic for Optimal Planning. 2007.

[3] Florian Pommerening, Gabriele Roger, Malte Helmert, Blai Bonet. LP-Based Heuristics for Cost-Optimal Planning. 2014. 

[4] Minh B. Do, Subbarao Kambhampati. Improving the temporal flexibility of position constrained metric temporal plans. 2003.

[5] Christian Muise, Sheila A. McIlraith, J. Christopher Beck. Optimally Relaxing Partial-Order Plans with MaxSAT. 2012.

[6] Buser Say, Andre A. Cire, J. Christopher Beck. Mathematical Programming Models For Optimizing Partial-Order Plan Flexibility. 2016.

[7] Toby Davies, Adrian R. Pearce, Peter Stuckey, Nir Lipovetzky. Sequencing Operator Counts. 2015.

Required knowledge

A successful candidate should have proficient programming skills (e.g., in Python) as well as background in at least one of the following:

  • automated planning, and/or
  • mathematical modeling.