Mostrar el registro sencillo del ítem

dc.contributor.advisorVelasco Gregory, Mauricio Fernando
dc.contributor.authorBetancourt Cardona, Nicolás
dc.description.abstractAn essential resource in the preservation of earth's biodiversity is keeping large natural areas protected. Unfortunately, sites of ecological interest are constantly threatened by illegal actors and the manpower allocated to monitor and safeguard these spaces is often insufficient. People in charge of taking care of those areas have to optimally allocate the patrolling resources in extensive tracts of land and are often in great disadvantage against the attackers. This problem has been previously studied in the literature and most of the work focuses on open spaces like the African Savannah or requires a discretization of the protected area. These approaches do not capture the reality of South American parks where, due to the density of the vegetation and the ruggedness of the terrain, traveling is done only over a limited collection of available trails (a graph). The problem addressed in this work is the design of near-optimal patrol schedules for rangers in such graphs. We illustrate our results in the trail map of Jamacoaque, a natural reserve in Ecuador. Our central result is that the framework of combinatorial multi-armed bandits is very well suited for this problem and provide theoretical guarantees as well as experimental simulations of our proposed route suggestion algorithms. Additionally, we explore other research directions such as coupling the route suggestion algorithm with additional information provided by an acoustic monitoring system.
dc.format.extent48 páginases_CO
dc.publisherUniversidad de los Andeses_CO
dc.titleGreen Security Games Along Trails
dc.typeTrabajo de grado - Maestríaes_CO
dc.publisher.programMaestría en Matemáticases_CO
dc.subject.keywordMulti-armed Bandits
dc.subject.keywordReinforcement learning
dc.subject.keywordCombinatorial optimization
dc.subject.keywordSubmodular functions
dc.subject.keywordDecision making under uncertainty
dc.subject.keywordNatural resource management
dc.publisher.facultyFacultad de Cienciases_CO
dc.publisher.departmentDepartamento de Matemáticases_CO
dc.contributor.juryDilkina, Bistra
dc.contributor.juryRiascos Villegas, Alvaro José
dc.description.degreenameMagíster en Matemáticases_CO
dc.identifier.instnameinstname:Universidad de los Andeses_CO
dc.identifier.reponamereponame:Repositorio Institucional Sénecaes_CO
dc.relation.referencesM. R. Appleton, A. Courtiol, L. Emerton, et al., Protected area personnel and ranger numbers are insufficient to deliver global expectations, Nature Sustainability, 2022.es_CO
dc.relation.referencesD. Kar, T. H. Nguyen, F. Fang, M. Brown, A. Sinha, M. Tambe, and A. X. Jiang, Trends and applications in stackelberg security games, Handbook of Dynamic Game Theory, pp. 1223¿1269, 2018.es_CO
dc.relation.referencesL. Xu, E. Bondi, F. Fang, A. Perrault, K. Wang, and M. Tambe, Dual-mandate pa- trols: Multi-armed bandits for green security, CoRR, vol. abs/2009.06560, 2020. arXiv: 2009.06560.es_CO
dc.relation.referencesP. Bauer, The circuit polytope: Facets, Mathematics of Operations Research, vol. 22, no. 1, pp. 110-145, 1997.es_CO
dc.relation.referencesM. Zhu, X. Zheng, Y. Wang, Y. Li, and Q. Liang, Adaptive portfolio by solving multi- armed bandit via thompson sampling, eng,, 2019es_CO
dc.relation.referencesS. S. Villar, J. Bowden, and J. Wason, Multi-armed bandit models for the optimal design of clinical trials: Benefits and challenges, eng, Statistical science, vol. 30, no. 2, pp. 199-215, 2015.es_CO
dc.relation.referencesK. Liu and Q. Zhao, Adaptive shortest-path routing under unknown and stochastically varying link states, eng,, 2012.es_CO
dc.relation.referencesA. Badanidiyuru, R. Kleinberg, and A. Slivkins, Bandits with knapsacks, eng, Journal of the ACM, vol. 65, no. 3, pp. 1-55, 2018.es_CO
dc.relation.referencesF. Trovò, S. Paladino, M. Restelli, and N. Gatti, Improving multi-armed bandit algo- rithms in online pricing settings, eng, International journal of approximate reasoning, vol. 98, pp. 196-235, 2018.es_CO
dc.relation.referencesS. Ontañón, Combinatorial multi-armed bandits for real-time strategy games, eng, The Journal of artificial intelligence research, vol. 58, pp. 665-702, 2017.es_CO
dc.relation.referencesP. Auer, N. Cesa-Bianchi, and P. Fischer, Finite-time analysis of the multiarmed bandit problem - machine learning, 2002.es_CO
dc.relation.referencesS. Boucheron, G. Lugosi, and P. Massart, 18Basic Inequalities, in Concentration In- equalities: A Nonasymptotic Theory of Independence, Oxford University Press, 2013, pp. 18-51. eprint: chapter-pdf/43899181/acprof-9780199535255-chapter-02.pdf.es_CO
dc.relation.referencesT. L. Lai, H. Robbins, et al., Asymptotically efficient adaptive allocation rules, Ad- vances in applied mathematics, vol. 6, no. 1, pp. 4-22, 1985.es_CO
dc.relation.referencesW. Chen, Y. Wang, and Y. Yuan, Combinatorial multi-armed bandit: General frame- work and applications, in Proceedings of the 30th International Conference on Machine Learning, S. Dasgupta and D. McAllester, Eds., ser. Proceedings of Machine Learning Research, Atlanta, Georgia, USA: PMLR, 2013, pp. 151-159.es_CO
dc.relation.referencesA. Benoit, A Guide to Algorithm Design : Paradigms, Methods, and Complexity Anal- ysis (Chapman & Hall/CRC applied algorithms and data structures series), eng, 1ses_CO
dc.relation.referencesG. L. Nemhauser, L. A. Wolsey, and M. L. Fisher, An analysis of approximations for maximizing submodular set functions-i, Mathematical Programming, vol. 14, no. 1, pp. 265-294, 1978.es_CO
dc.relation.referencesS.-H. Lee, K.-J. Han, K. Lee, K.-J. Lee, K.-Y. Oh, and M.-J. Lee, Classification of landscape affected by deforestation using high-resolution remote sensing data and deep-learning techniques, eng, Remote sensing (Basel, Switzerland), vol. 12, no. 20, pp. 3372, 2020.es_CO
dc.relation.referencesY. E. Shimabukuro, Spectral Mixture for Remote Sensing Linear Model and Applica- tions (Springer Remote Sensing/Photogrammetry), eng, 1st ed. 2019. Cham: Springer International Publishing, 2019.es_CO
dc.relation.referencesS. Sagawa, P. W. Koh, T. Lee, et al., Extending the WILDS benchmark for unsuper- vised adaptation, CoRR, vol. abs/2112.05090, 2021. arXiv: 2112.05090.es_CO
dc.relation.referencesP. Kulits, J. Wall, A. Bedetti, M. Henley, and S. Beery, Elephantbook: A semi- automated human-in-the-loop system for elephant re-identification, CoRR, vol. abs/2106.15083, 2021. arXiv: 2106.15083.es_CO
dc.relation.referencesM. S. Norouzzadeh, D. Morris, S. Beery, N. Joshi, N. Jojic, and J. Clune, A deep active learning system for species identification and counting in camera trap images, CoRR, vol. abs/1910.09716, 2019. arXiv: 1910.09716.es_CO
dc.relation.referencesD. Tuia, B. Kellenberger, S. Beery, et al., Perspectives in machine learning for wildlife conservation, Nature Communications, vol. 13, no. 1, 2022.es_CO

Ficheros en el ítem


Nombre: trabajo_grado_nicolas_betancou ...

Este ítem aparece en la(s) siguiente(s) colección(ones)

Mostrar el registro sencillo del ítem