Enhancing Operational Efficiency in Water Canals Using Sequential Fuzzy Reinforcement Learning
Shahverdi, Kazem ; Loni, Reyhaneh ; Yoosefdoost, Icen ; Berndtsson, Ronny LU
; Kaviani, Mitra
and Yoosefdoost, Arash
(2026)
In Water Resources Management
40.
- Abstract
Pressurized irrigation systems on farms are supplied by irrigation canals, whose poor performance results in inadequate performance of the pressurized irrigation systems. The potential role of artificial intelligence (AI) in industry and agriculture, as well as in irrigation canals, could be substantial. For a part of an irrigation canal’s operation (small scale), it is possible to see some examples of applied AI techniques. To investigate the performance of AI and its potential applications in large and medium-sized irrigation canal systems, a sequential fuzzy-based reinforcement learning model was developed, formulated, and implemented in this study using the SARSA process, which learns based on the sequential performance of action,... (More)
Pressurized irrigation systems on farms are supplied by irrigation canals, whose poor performance results in inadequate performance of the pressurized irrigation systems. The potential role of artificial intelligence (AI) in industry and agriculture, as well as in irrigation canals, could be substantial. For a part of an irrigation canal’s operation (small scale), it is possible to see some examples of applied AI techniques. To investigate the performance of AI and its potential applications in large and medium-sized irrigation canal systems, a sequential fuzzy-based reinforcement learning model was developed, formulated, and implemented in this study using the SARSA process, which learns based on the sequential performance of action, state, reward, action, state. The developed model has three main modules: fuzzy, SARSA, and simulator modules. To make the simulator adaptable, an improved simulator was proposed. The model was evaluated for the Aghili East Canal in Iran. The findings were satisfactory, and the efficiency and adequacy indicators were found to be 0.943 and 0.922, respectively. As a result, there are potential applications for similar irrigation systems to support sustainable and efficient water management.
(Less)
- author
- Shahverdi, Kazem
; Loni, Reyhaneh
; Yoosefdoost, Icen
; Berndtsson, Ronny
LU
; Kaviani, Mitra
and Yoosefdoost, Arash
- organization
- publishing date
- 2026-03-14
- type
- Contribution to journal
- publication status
- published
- subject
- keywords
- Fuzzy system, Irrigation canals, Reinforcement learning, Water distribution, Water management
- in
- Water Resources Management
- volume
- 40
- article number
- 189
- publisher
- Springer Science and Business Media B.V.
- external identifiers
-
- scopus:105033804567
- ISSN
- 0920-4741
- DOI
- 10.1007/s11269-026-04554-x
- language
- English
- LU publication?
- yes
- additional info
- Publisher Copyright: © The Author(s), under exclusive licence to Springer Nature B.V. 2026.
- id
- 67d64099-d360-4565-9255-9cf5c246b83c
- date added to LUP
- 2026-04-14 08:41:20
- date last changed
- 2026-05-21 15:25:41
@article{67d64099-d360-4565-9255-9cf5c246b83c,
abstract = {{<p>Pressurized irrigation systems on farms are supplied by irrigation canals, whose poor performance results in inadequate performance of the pressurized irrigation systems. The potential role of artificial intelligence (AI) in industry and agriculture, as well as in irrigation canals, could be substantial. For a part of an irrigation canal’s operation (small scale), it is possible to see some examples of applied AI techniques. To investigate the performance of AI and its potential applications in large and medium-sized irrigation canal systems, a sequential fuzzy-based reinforcement learning model was developed, formulated, and implemented in this study using the SARSA process, which learns based on the sequential performance of action, state, reward, action, state. The developed model has three main modules: fuzzy, SARSA, and simulator modules. To make the simulator adaptable, an improved simulator was proposed. The model was evaluated for the Aghili East Canal in Iran. The findings were satisfactory, and the efficiency and adequacy indicators were found to be 0.943 and 0.922, respectively. As a result, there are potential applications for similar irrigation systems to support sustainable and efficient water management.</p>}},
author = {{Shahverdi, Kazem and Loni, Reyhaneh and Yoosefdoost, Icen and Berndtsson, Ronny and Kaviani, Mitra and Yoosefdoost, Arash}},
issn = {{0920-4741}},
keywords = {{Fuzzy system; Irrigation canals; Reinforcement learning; Water distribution; Water management}},
language = {{eng}},
month = {{03}},
publisher = {{Springer Science and Business Media B.V.}},
series = {{Water Resources Management}},
title = {{Enhancing Operational Efficiency in Water Canals Using Sequential Fuzzy Reinforcement Learning}},
url = {{http://dx.doi.org/10.1007/s11269-026-04554-x}},
doi = {{10.1007/s11269-026-04554-x}},
volume = {{40}},
year = {{2026}},
}