Conditional Epistemic Planning | Mads Krogsgaard Pico Jensen
| | Abstract | Dynamic epistemic logic reasons about higher-order knowledge in multi-agent systems, and has shown promise as a framework for use in automated planning. Using dynamic epistemic logic in automated planning allows for capturing of partial-observability and non-determinism, making it an highly expressive combination. This thesis is based on the later work within this field by Andersen, Bolander, and Jensen [1], wherein a planning algorithm is proposed for finding strong and weak plans in a single-agent setting. Their planning algorithm takes the internal view of the agent, necessary for developing planners for use in autonomous agents.
The main contribution made by this thesis is a planning algorithm for finding strong cyclic and weak plans in a multi-agent setting with one acting agent using dynamic epistemic logic. The algorithm takes the internal perspective of the acting agent, thereby making it the planning agent. The planning algorithm
is shown to be both sound and complete, and the work is completed by an implementation of the proposed planning algorithm. | | Type | Master's thesis [Academic thesis] | | Year | 2013 | | Publisher | Technical University of Denmark, DTU Compute, E-mail: compute@compute.dtu.dk | | Address | Matematiktorvet, Building 303-B, DK-2800 Kgs. Lyngby, Denmark | | Series | M.Sc.-2013-27 | | Note | DTU supervisor: Thomas Bolander, tobo@dtu.dk, DTU Compute | | Electronic version(s) | [pdf] | | Publication link | http://www.compute.dtu.dk/English.aspx | | BibTeX data | [bibtex] | | IMM Group(s) | Computer Science & Engineering |
|