{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,8,21]],"date-time":"2026-08-21T18:24:25Z","timestamp":1787336665648,"version":"build-2736575974"},"reference-count":30,"publisher":"Society for Industrial & Applied Mathematics (SIAM)","issue":"2","content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["SIAM J. Sci. Comput."],"published-print":{"date-parts":[[2023,4,30]]},"abstract":"<jats:p>Abstract.<\/jats:p>\n                  <jats:p>Many problems in science and engineering require an efficient numerical approximation of integrals or solutions to differential equations. For systems with rapidly changing dynamics, an equidistant discretization is often inadvisable as it results in prohibitively large errors or computational effort. To this end, adaptive schemes, such as solvers based on Runge\u2013Kutta pairs, have been developed which adapt the step size based on local error estimations at each step. While the classical schemes apply very generally and are highly efficient on regular systems, they can behave suboptimally when an inefficient step rejection mechanism is triggered by structurally complex systems such as chaotic systems. To overcome these issues, we propose a method to tailor numerical schemes to the problem class at hand. This is achieved by combining simple, classical quadrature rules or ODE solvers with data-driven time-stepping controllers. Compared with learning solution operators to ODEs directly, it generalizes better to unseen initial data as our approach employs classical numerical schemes as base methods. At the same time it can make use of identified structures of a problem class and, therefore, outperforms state-of-the-art adaptive schemes. Several examples demonstrate superior efficiency. Source code is available at https:\/\/github.com\/lueckem\/quadrature-ML .<\/jats:p>","DOI":"10.1137\/21m1412682","type":"journal-article","created":{"date-parts":[[2023,4,26]],"date-time":"2023-04-26T03:39:20Z","timestamp":1682480360000},"page":"A579-A595","source":"Crossref","is-referenced-by-count":8,"title":["Efficient Time-Stepping for Numerical Integration Using Reinforcement Learning"],"prefix":"10.1137","volume":"45","author":[{"given":"Michael","family":"Dellnitz","sequence":"first","affiliation":[{"name":"Department of Mathematics, Paderborn University, Paderborn, Germany."}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Eyke","family":"H\u00fcllermeier","sequence":"additional","affiliation":[{"name":"Department of Computer Science, LMU Munich, Munich, Germany."}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-8415-0038","authenticated-orcid":true,"given":"Marvin","family":"L\u00fccke","sequence":"additional","affiliation":[{"name":"Modeling and Simulation of Complex Processes, Zuse Institute Berlin, Berlin, Germany."}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Sina","family":"Ober-Bl\u00f6baum","sequence":"additional","affiliation":[{"name":"Department of Mathematics, Paderborn University, Paderborn, Germany."}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-5940-8057","authenticated-orcid":true,"given":"Christian","family":"Offen","sequence":"additional","affiliation":[{"name":"Department of Mathematics, Paderborn University, Paderborn, Germany."}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-3389-793X","authenticated-orcid":true,"given":"Sebastian","family":"Peitz","sequence":"additional","affiliation":[{"name":"Department of Computer Science, Paderborn University, Paderborn, Germany."}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Karlson","family":"Pfannschmidt","sequence":"additional","affiliation":[{"name":"Department of Computer Science, Paderborn University, Paderborn, Germany."}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"351","published-online":{"date-parts":[[2023,4,26]]},"reference":[{"key":"ref1","unstructured":"Scipy v1.5.4 Reference Guide: Integration and ODEs (scipy.integrate.quad), https:\/\/docs.scipy.org\/doc\/scipy\/reference\/generated\/scipy.integrate.quad.html."},{"key":"ref2","series-title":"Texts Appl. Math. 42","volume-title":"Scientific Computing with Ordinary Differential Equations","author":"Deuflhard P.","year":"2012"},{"key":"ref3","doi-asserted-by":"publisher","DOI":"10.1007\/978-0-387-21584-6"},{"key":"ref4","doi-asserted-by":"publisher","DOI":"10.1016\/0771-050X(80)90013-3"},{"key":"ref5","unstructured":"J. Fan , \nZ. Wang , \nY. Xie , and \nZ. Yang , A theoretical analysis of deep Q-learning, in Proceedings of the 2nd Conference on Learning for Dynamics and Control, \nA. M. Bayen , \nA. Jadbabaie , \nG. Pappas , \nP. A. Parrilo , \nB. Recht , \nC. Tomlin , and \nM. Zeilinger , eds. Proc. Mach. Learn. Res. 120, 2020, pp. 486\u2013489."},{"key":"ref6","unstructured":"X. Glorot  and \nY. Bengio , Understanding the difficulty of training deep feedforward neural networks, in Proceedings of the Thirteenth International Conference on Artificial Intelligence and Statistics, \nY. W. Teh  and \nM. Titterington , eds. Proc. Mach. Learn. Res. 9, PMLR, 2010, pp. 249\u2013256, http:\/\/proceedings.mlr.press\/v9\/glorot10a.html."},{"key":"ref7","series-title":"Springer Ser. Comput. Math. 31","volume-title":"Geometric Numerical Integration: Structure-Preserving Algorithms for Ordinary Differential Equations","author":"Hairer E.","year":"2013"},{"key":"ref8","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-642-05221-7"},{"key":"ref9","volume-title":"Solving Ordinary Differential Equations I","author":"Hairer E.","year":"1993"},{"key":"ref10","unstructured":"M. J. Kearns  and \nS. P. Singh , Finite-sample convergence rates for Q-learning and indirect algorithms, in Advances in Neural Information Processing Systems, 1999, pp. 996\u20131002."},{"key":"ref11","unstructured":"D. P. Kingma  and \nJ. Ba , Adam: A Method for Stochastic Optimization, https:\/\/arxiv.org\/abs\/1412.6980, 2014."},{"key":"ref12","doi-asserted-by":"publisher","DOI":"10.1016\/j.physd.2020.132416"},{"key":"ref13","doi-asserted-by":"publisher","DOI":"10.1137\/18M1198375"},{"key":"ref14","unstructured":"T. P. Lillicrap , \nJ. J. Hunt , \nA. Pritzel , \nN. Heess , \nT. Erez , \nY. Tassa , \nD. Silver , and \nD. Wierstra , Continuous Control with Deep Reinforcement Learning, poster, International Conference on Learning Representations, 2016."},{"key":"ref15","doi-asserted-by":"publisher","DOI":"10.1098\/rsta.2021.0200"},{"key":"ref16","doi-asserted-by":"publisher","DOI":"10.1038\/nature14236"},{"key":"ref17","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-642-61786-7"},{"key":"ref18","doi-asserted-by":"publisher","DOI":"10.1103\/PhysRevD.100.016002"},{"key":"ref19","doi-asserted-by":"publisher","DOI":"10.1016\/j.jcp.2018.10.045"},{"key":"ref20","doi-asserted-by":"publisher","DOI":"10.1016\/j.jcp.2019.07.050"},{"key":"ref21","unstructured":"SciPy, v1.5.4 Reference Guide: Integration and ODEs (scipy.integrate.solve_ivp), https:\/\/docs.scipy.org\/doc\/scipy\/reference\/generated\/scipy.integrate.solve_ivp.html."},{"key":"ref22","doi-asserted-by":"publisher","DOI":"10.1016\/j.cam.2006.11.021"},{"key":"ref23","doi-asserted-by":"publisher","DOI":"10.1137\/S1064827594276424"},{"key":"ref24","doi-asserted-by":"publisher","DOI":"10.1119\/1.16860"},{"key":"ref25","doi-asserted-by":"publisher","DOI":"10.1016\/j.jcp.2018.08.029"},{"key":"ref26","volume-title":"Reinforcement Learning: An Introduction","author":"Sutton R. S.","year":"2018"},{"key":"ref27","unstructured":"M. Tsatsos , The Van der Pol Equation, Ph.D. thesis, https:\/\/arxiv.org\/abs\/0803.1658, 2008."},{"key":"ref28","doi-asserted-by":"publisher","DOI":"10.1038\/s41592-019-0686-2"},{"key":"ref29","doi-asserted-by":"publisher","DOI":"10.1007\/BF00992698"},{"key":"ref30","unstructured":"Wolfram Language & System Documentation Center: The Design of the NDSolve Framework, Wolfram Research, https:\/\/reference.wolfram.com\/language\/tutorial\/NDSolveDesign.html."}],"container-title":["SIAM Journal on Scientific Computing"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/epubs.siam.org\/doi\/pdf\/10.1137\/21M1412682","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,8,21]],"date-time":"2026-08-21T17:25:48Z","timestamp":1787333148000},"score":1,"resource":{"primary":{"URL":"https:\/\/epubs.siam.org\/doi\/10.1137\/21M1412682"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2023,4,26]]},"references-count":30,"journal-issue":{"issue":"2","published-print":{"date-parts":[[2023,4,30]]}},"alternative-id":["10.1137\/21M1412682"],"URL":"https:\/\/doi.org\/10.1137\/21m1412682","relation":{},"ISSN":["1064-8275","1095-7197"],"issn-type":[{"value":"1064-8275","type":"print"},{"value":"1095-7197","type":"electronic"}],"subject":[],"published":{"date-parts":[[2023,4,26]]}}}