Monte mario: Platforming with MCTS

Emil Juul Jacobsen; Rasmus Greve; Julian Togelius

doi:10.1145/2576768.2598392

Monte mario: Platforming with MCTS

Emil Juul Jacobsen, Rasmus Greve, Julian Togelius

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Abstract

Monte Carlo Tree Search (MCTS) is applied to control the player character in a clone of the popular platform game Su- per Mario Bros. Standard MCTS is applied through search in state space with the goal of moving the furthest to the right as quickly as possible. Despite parameter tuning, only moderate success is reached. Several modifications to the algorithm are then introduced specifically to deal with the behavioural pathologies that were observed. Two of the modifications are to our best knowledge novel. A combination of these modifications is found to lead to almost perfect play on linear levels. Furthermore, when adding noise to the benchmark, MCTS outperforms the best known algorithm for these levels. The analysis and algorithmic innovations in this paper are likely to be useful when applying MCTS to other video games.

Original language	English (US)
Title of host publication	GECCO 2014 - Proceedings of the 2014 Genetic and Evolutionary Computation Conference
Publisher	Association for Computing Machinery
Pages	293-300
Number of pages	8
ISBN (Print)	9781450326629
DOIs	https://doi.org/10.1145/2576768.2598392
State	Published - 2014
Event	16th Genetic and Evolutionary Computation Conference, GECCO 2014 - Vancouver, BC, Canada Duration: Jul 12 2014 → Jul 16 2014

Publication series

Name	GECCO 2014 - Proceedings of the 2014 Genetic and Evolutionary Computation Conference

Other

Other	16th Genetic and Evolutionary Computation Conference, GECCO 2014
Country	Canada
City	Vancouver, BC
Period	7/12/14 → 7/16/14

ASJC Scopus subject areas

Computational Theory and Mathematics
Applied Mathematics

Access to Document

10.1145/2576768.2598392

Fingerprint Dive into the research topics of 'Monte mario: Platforming with MCTS'. Together they form a unique fingerprint.

Cite this

Jacobsen, E. J., Greve, R., & Togelius, J. (2014). Monte mario: Platforming with MCTS. In GECCO 2014 - Proceedings of the 2014 Genetic and Evolutionary Computation Conference (pp. 293-300). (GECCO 2014 - Proceedings of the 2014 Genetic and Evolutionary Computation Conference). Association for Computing Machinery. https://doi.org/10.1145/2576768.2598392

Monte mario : Platforming with MCTS. / Jacobsen, Emil Juul; Greve, Rasmus; Togelius, Julian.

GECCO 2014 - Proceedings of the 2014 Genetic and Evolutionary Computation Conference. Association for Computing Machinery, 2014. p. 293-300 (GECCO 2014 - Proceedings of the 2014 Genetic and Evolutionary Computation Conference).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Jacobsen, EJ, Greve, R & Togelius, J 2014, Monte mario: Platforming with MCTS. in GECCO 2014 - Proceedings of the 2014 Genetic and Evolutionary Computation Conference. GECCO 2014 - Proceedings of the 2014 Genetic and Evolutionary Computation Conference, Association for Computing Machinery, pp. 293-300, 16th Genetic and Evolutionary Computation Conference, GECCO 2014, Vancouver, BC, Canada, 7/12/14. https://doi.org/10.1145/2576768.2598392

@inproceedings{b4cea88935864b87911b4121e118241c,

title = "Monte mario: Platforming with MCTS",

abstract = "Monte Carlo Tree Search (MCTS) is applied to control the player character in a clone of the popular platform game Su- per Mario Bros. Standard MCTS is applied through search in state space with the goal of moving the furthest to the right as quickly as possible. Despite parameter tuning, only moderate success is reached. Several modifications to the algorithm are then introduced specifically to deal with the behavioural pathologies that were observed. Two of the modifications are to our best knowledge novel. A combination of these modifications is found to lead to almost perfect play on linear levels. Furthermore, when adding noise to the benchmark, MCTS outperforms the best known algorithm for these levels. The analysis and algorithmic innovations in this paper are likely to be useful when applying MCTS to other video games.",

author = "Jacobsen, {Emil Juul} and Rasmus Greve and Julian Togelius",

year = "2014",

doi = "10.1145/2576768.2598392",

language = "English (US)",

isbn = "9781450326629",

series = "GECCO 2014 - Proceedings of the 2014 Genetic and Evolutionary Computation Conference",

publisher = "Association for Computing Machinery",

pages = "293--300",

booktitle = "GECCO 2014 - Proceedings of the 2014 Genetic and Evolutionary Computation Conference",

note = "16th Genetic and Evolutionary Computation Conference, GECCO 2014 ; Conference date: 12-07-2014 Through 16-07-2014",

}

TY - GEN

T1 - Monte mario

T2 - 16th Genetic and Evolutionary Computation Conference, GECCO 2014

AU - Jacobsen, Emil Juul

AU - Greve, Rasmus

AU - Togelius, Julian

PY - 2014

Y1 - 2014

N2 - Monte Carlo Tree Search (MCTS) is applied to control the player character in a clone of the popular platform game Su- per Mario Bros. Standard MCTS is applied through search in state space with the goal of moving the furthest to the right as quickly as possible. Despite parameter tuning, only moderate success is reached. Several modifications to the algorithm are then introduced specifically to deal with the behavioural pathologies that were observed. Two of the modifications are to our best knowledge novel. A combination of these modifications is found to lead to almost perfect play on linear levels. Furthermore, when adding noise to the benchmark, MCTS outperforms the best known algorithm for these levels. The analysis and algorithmic innovations in this paper are likely to be useful when applying MCTS to other video games.

AB - Monte Carlo Tree Search (MCTS) is applied to control the player character in a clone of the popular platform game Su- per Mario Bros. Standard MCTS is applied through search in state space with the goal of moving the furthest to the right as quickly as possible. Despite parameter tuning, only moderate success is reached. Several modifications to the algorithm are then introduced specifically to deal with the behavioural pathologies that were observed. Two of the modifications are to our best knowledge novel. A combination of these modifications is found to lead to almost perfect play on linear levels. Furthermore, when adding noise to the benchmark, MCTS outperforms the best known algorithm for these levels. The analysis and algorithmic innovations in this paper are likely to be useful when applying MCTS to other video games.

UR - http://www.scopus.com/inward/record.url?scp=84905715070&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84905715070&partnerID=8YFLogxK

U2 - 10.1145/2576768.2598392

DO - 10.1145/2576768.2598392

M3 - Conference contribution

AN - SCOPUS:84905715070

SN - 9781450326629

T3 - GECCO 2014 - Proceedings of the 2014 Genetic and Evolutionary Computation Conference

SP - 293

EP - 300

BT - GECCO 2014 - Proceedings of the 2014 Genetic and Evolutionary Computation Conference

PB - Association for Computing Machinery

Y2 - 12 July 2014 through 16 July 2014

ER -