{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/sgd-for-robot-motion-the-effectiveness-of","title":"SGD for robot motion? The effectiveness of stochastic optimization on a new benchmark for biped locomotion tasks","arxiv_id":"1710.03029","date":"2017-10-09","proceeding":null,"authors":["Martim Brandao","Kenji Hashimoto","Atsuo Takanishi"],"abstract":"Trajectory optimization and posture generation are hard problems in robot\nlocomotion, which can be non-convex and have multiple local optima. Progress on\nthese problems is further hindered by a lack of open benchmarks, since\ncomparisons of different solutions are difficult to make. In this paper we\nintroduce a new benchmark for trajectory optimization and posture generation of\nlegged robots, using a pre-defined scenario, robot and constraints, as well as\nevaluation criteria. We evaluate state-of-the-art trajectory optimization\nalgorithms based on sequential quadratic programming (SQP) on the benchmark, as\nwell as new stochastic and incremental optimization methods borrowed from the\nlarge-scale machine learning literature. Interestingly we show that some of\nthese stochastic and incremental methods, which are based on stochastic\ngradient descent (SGD), achieve higher success rates than SQP on tough\ninitializations. Inspired by this observation we also propose a new incremental\nvariant of SQP which updates only a random subset of the costs and constraints\nat each iteration. The algorithm is the best performing in both success rate\nand convergence speed, improving over SQP by up to 30% in both criteria. The\nbenchmark's resources and a solution evaluation script are made openly\navailable.","url_abs":"http://arxiv.org/abs/1710.03029v1","url_pdf":"http://arxiv.org/pdf/1710.03029v1.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"sgd-for-robot-motion-the-effectiveness-of","repo_url":"https://github.com/martimbrandao/legopt-benchmark","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"none","reach":null}],"tasks":[{"task_slug":"stochastic-optimization","task_name":"Stochastic Optimization"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":null,"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}