This archive contains the manuscript PDF and LaTeX source for “Reaction-State Adaptive Transmutation Framework (RATF).” The work introduces a conceptual framework for modeling nuclear transmutation as a partially observed, feedback-controlled reaction network. RATF combines latent reaction-state estimation, softmax-based pathway occupancy modeling, and adaptive external actuation to study how reaction pathways can be steered toward favorable transmutation outcomes. The manuscript includes an illustrative three-isotope toy simulation comparing uncontrolled dynamics, static control, and RATF adaptive control. In this simulation, RATF reduces the final high-risk isotope fraction, increases stable-product formation, increases mean target-pathway occupancy, and slightly lowers total cost relative to the static baseline. This study is intended as a methodological and conceptual demonstration. The numerical parameters are illustrative and are not calibrated to a specific physical reactor or transmutation system. This slim archive contains only the final manuscript PDF and the LaTeX source file. External figure assets are not included, so the PDF should be treated as the canonical reading version.
Hakjun Kim (Thu,) studied this question.