Javascript must be enabled for the correct page display

THEIA: A labeling based backtracking solver for Abstract Argumentation

Kinder, Lukas (2021) THEIA: A labeling based backtracking solver for Abstract Argumentation. Bachelor's Thesis, Artificial Intelligence.

[img]
Preview
Text
THEIA_Bachelor_Thesis_Lukas_Kinder_s3686566.pdf

Download (621kB) | Preview
[img] Text
toestemming.pdf
Restricted to Registered users only

Download (116kB)

Abstract

THEIA is a labeling based algorithm to find complete sets of a Dung argumentation framework. State of the art backtracking solvers do this by repeatedly choosing an argument and label it until either a contradiction with respect to the labels is reached or a solution is found. The main idea of THEIA is to reduce the number of backtracking steps by using propagation techniques that keep track of arguments that cannot be defeated or undefeated. To assess the performance, the program was tested on the data-set of the ICCMA 2019 and was in general faster than the backtracking solvers HEUREKA and DREDD. This result shows that backtracking solvers can be improved by using a bigger set of labels which enable more powerful propagation techniques.

Item Type: Thesis (Bachelor's Thesis)
Supervisor name: Verheij, H.B.
Degree programme: Artificial Intelligence
Thesis type: Bachelor's Thesis
Language: English
Date Deposited: 06 Aug 2021 06:04
Last Modified: 24 Aug 2021 09:10
URI: https://fse.studenttheses.ub.rug.nl/id/eprint/25428

Actions (login required)

View Item View Item