Supervisory gate control of the South Mirzachul main canal under supply shortage: code, data and results
Annotatsiya
Code, input data, trained controllers and simulation results for the article "Neural supervisory gate control for sharing a supply shortage in a large irrigation main canal". The study tests whether a neural supervisory controller, trained by an evolution strategy on a one-dimensional Saint-Venant model of the 123.7-km South Mirzachul main canal (Uzbekistan), shares a supply shortage more equitably than a fixed rule and than a transparent rule optimized with the same algorithm and budget, using only the gates that the canal already has. It also tests the controllers under roughness uncertainty, measurement noise and equipment faults. The repository contains everything needed to rebuild Tables 1–6, Table B.1 and Figs. 2–9 of the article. Code: MIT licence; data, results and figures: CC BY 4.0 (see LICENSE-DATA.md).
Hali tarjima qilinmagan