[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"branding":3,"analytics":7,"article-rl-tuned-solver-cuts-linear-programming-iterations-up-to-56x":10,"sections":35},{"siteName":4,"siteTagline":5,"publisherName":4,"contactEmail":6},"The Revision","Tech news, decoded.","editor@therevision.news",{"gaMeasurementId":8,"adsenseClientId":9},"G-ZW2MV82GYR","ca-pub-8533917693782264",{"article":11},{"id":12,"slug":13,"title":14,"dek":15,"body_md":16,"tags_json":17,"published_at":18,"created_at":19,"updated_at":20,"status":21,"review_note":22,"review_notes":23,"image_url":22,"persona_id":22,"persona_name":22,"section":24,"tags":25,"sources":30,"feedback":34,"feedback_at":22,"cost_usd":34,"total_tokens":34},9738,"rl-tuned-solver-cuts-linear-programming-iterations-up-to-56x","RL-Tuned Solver Cuts Linear Programming Iterations Up to 5.6x","A reinforcement-learning policy tunes a GPU solver's parameters and restarts, cutting iterations and runtime on linear programs with millions of variables.","A new reinforcement-learning policy teaches a GPU linear-programming solver when to accelerate and when to stop and restart.\n\nResearchers introduced GALLOP, a method that trains an RL policy to tune the primal-dual hybrid gradient algorithm used to solve large linear programs on GPUs. Rather than hand-tuning acceleration parameters and restart timing, GALLOP learns both using a modified proximal policy optimization objective that treats different control groups separately. Tested on six linear-programming families plus a public item-placement benchmark, it cut iteration counts by 1.9x to 5.6x and sped up wall-clock time by as much as 16x compared with MPAX, an existing GPU solver. A single trained policy generalized to problems 3x to 400x larger than its training data, including a transport problem with 10.24 million variables.\n\nLinear programming runs the unglamorous machinery behind supply-chain routing, ad allocation, and resource scheduling, and the bottleneck lately has been tuning, not theory. Hand-configuring solvers for each new problem type is tedious work that does not scale well. A learned policy that generalizes far beyond its training size hints that reinforcement learning could become a standard part of numerical solvers, not just a trick reserved for game-playing systems and chatbots.\n\nThe gains are measured against MPAX, a GPU-native solver. Whether they carry over to the commercial solvers most companies currently run is still untested.","[\"reinforcement-learning\",\"linear-programming\",\"gpu-computing\",\"optimization\"]","2026-10-02T04:00:00.000Z","2026-10-03T08:31:24.070Z","2026-10-03T08:31:29.272Z","published",null,[],"ai",[26,27,28,29],"reinforcement-learning","linear-programming","gpu-computing","optimization",[31],{"name":32,"url":33},"arXiv cs.AI","https:\u002F\u002Farxiv.org\u002Fabs\u002F2610.01546",0,{"sections":36},[37,40,44,48,53,57,61,66,71,76,81,86,91,96],{"name":38,"slug":24,"count":39,"latest_published_at":18},"AI",6041,{"name":41,"slug":42,"count":43,"latest_published_at":18},"Security","security",848,{"name":45,"slug":46,"count":47,"latest_published_at":18},"Policy","policy",439,{"name":49,"slug":50,"count":51,"latest_published_at":52},"Deals","deals",317,"2026-10-01T22:00:00.000Z",{"name":54,"slug":55,"count":56,"latest_published_at":18},"Hardware","hardware",199,{"name":58,"slug":59,"count":60,"latest_published_at":18},"Science","science",176,{"name":62,"slug":63,"count":64,"latest_published_at":65},"Consumer Tech","consumer-tech",155,"2026-10-01T19:54:10.000Z",{"name":67,"slug":68,"count":69,"latest_published_at":70},"Dev Tools","dev-tools",96,"2026-10-01T16:57:03.000Z",{"name":72,"slug":73,"count":74,"latest_published_at":75},"Software","software",93,"2026-09-30T21:41:11.000Z",{"name":77,"slug":78,"count":79,"latest_published_at":80},"Startups","startups",90,"2026-10-01T21:55:22.000Z",{"name":82,"slug":83,"count":84,"latest_published_at":85},"Gaming","gaming",53,"2026-10-02T02:50:39.000Z",{"name":87,"slug":88,"count":89,"latest_published_at":90},"General","general",50,"2026-09-30T21:37:54.000Z",{"name":92,"slug":93,"count":94,"latest_published_at":95},"Reviews","reviews",31,"2026-09-28T14:31:34.000Z",{"name":97,"slug":98,"count":99,"latest_published_at":100},"How-To","how-to",7,"2026-10-01T09:00:00.000Z"]