Multi-Agent Programming in Jason

Rasmus Jensen, Bastian Buch

AbstractThe goal of the thesis is to learn the Jason programming language and use our learned knowledge to successfully analyse and evaluate the strategies used by the UFSCTeam2013 Jason code in relation to the Multi-agent contest scenario in 2013. Once we have analysed both the UFSCTeam2013's program and the scenario we will discus possible additions to the program that would increase its performance. These improvements will then be implemented in our own version which we call JABARA.

To either prove or disprove the effectiveness of our improvements we will run a series of tests to determine whether or not our implementations has improved the UFSCTeam2013's program. These tests will be simulations where our version competes against the UFSCTeam2013's version.

Our results show that we have indeed managed to improve the program, albeit only slightly. Our improved version deals with some scenarios better than UFSCTeam2013. Specifically, in maps with a high number of edges our agents survey more edges and gain more achievement points faster as a result.
TypeBachelor thesis [Academic thesis]
Year2014
PublisherTechnical University of Denmark, Department of Applied Mathematics and Computer Science
AddressRichard Petersens Plads, Building 324, DK-2800 Kgs. Lyngby, Denmark, compute@compute.dtu.dk
SeriesDTU Compute B.Sc.-2014
Note
Electronic version(s)[pdf]
Publication linkhttp://www.compute.dtu.dk/English.aspx
BibTeX data [bibtex]
IMM Group(s)Computer Science & Engineering