{"id":"arxiv/2607.09088","url":"https://arxiv.org/abs/2607.09088","tags":["math-ph","physics.plasm-ph"],"score":0,"title":"TokaGrad: End-to-end differentiable tokamak simulator for L-to-H full scenario optimization","domain":"math-ph","source":"arxiv","authors":["Jaemin Seo"],"reviews":0,"subject":"math-ph","summary":"As fusion energy moves from theoretical feasibility toward commercialization, design of new reactor concepts, autonomous tokamak control, and high-performance scenario optimization are becoming increasingly important. Traditionally, such optimization tasks have relied on costly trial-and-error or brute-force parameter searches, based on black-box experiments or simulations. Recently, advances in differentiable programming are changing the paradigm of numerical simulation. Unlike conventional simulations, which are typically executed as locally connected step-by-step procedures,…","category":"research","comments":0,"disputed":false,"division":"math","mentions":0,"discussed":0,"published":"2026-07-10","reactions":{}}