Hybrid Reinforcement Learning (RL) and Social Network Community Detection Framework for Optimal Management and Redesign of Urban Water Network

Sajjad Ali  Ettyem; Hasan Ziafat; Jafaar  Fahad A.Rida; Sanaz  Asadini

Authors

Sajjad Ali Ettyem Department of Computer Engineering, Isf.C., Islamic Azad University, Isfahan, Iran Author https://orcid.org/0009-0001-4247-527X
Hasan Ziafat Department of computer engineering, Kas.C., Islamic Azad University, Kashan, Iran https://orcid.org/0000-0001-6078-8088
Jafaar Fahad A.Rida Department of Communication Engineering, College of Engineering, University of Sumer, Dhiqar, Iraq Author https://orcid.org/0009-0007-7442-0314
Sanaz Asadini Department of Computer Engineering, Isf.C., Islamic Azad University, Isfahan, Iran Author https://orcid.org/0009-0004-7355-1206

Keywords:

Water Network Sectorization, District Metered Areas (DMAs), Community Detection, Reinforcement Learning, Multi-Objective Optimization

Abstract

This study presents a novel framework for water network sectorization that integrates reinforcement learning (Proximal Policy Optimization), social network search algorithms, and multi-objective optimization to simultaneously address water quality monitoring, hydraulic performance, and topological efficiency. Tested on the Thi-Qar, Iraq water distribution network, the proposed approach achieved substantial improvements: resilience deviation index dropped by 70%, balance index reached near-perfect uniformity (1.01), modularity improved by 9.7%, and average path length shortened by 11.9% while its deviation decreased by 43.5%. Critically, the framework fills a key literature gap by explicitly incorporating water quality objectives—minimizing detection time and contaminated volume while maximizing detection likelihood—alongside traditional energy and topological metrics. Results demonstrate that reinforcement learning-enhanced optimization effectively balances multiple competing objectives, offering water utilities a powerful decision-support tool for designing resilient, monitorable, and operationally efficient distribution systems, particularly valuable for developing regions where infrastructure investments require optimal allocation.

References

Arif, T. (2015). The Mathematics of Social Network Analysis: Metrics for Academic Social Networks. International Journal of Computer Applications Technology and Research, 4(12), 889-893. https://doi.org/10.7753/IJCATR0412.1003

Ayall, T. A., Liu, H., Zhou, C., Seid, A. M., Gereme, F. B., Abishu, H. N., & Yacob, Y. H. (2022). Graph Computing Systems and Partitioning Techniques: A Survey. IEEE Access, 10, 118523-118550. https://doi.org/10.1109/ACCESS.2022.3219422

Bahrami Chegeni, I., Riyahi, M. M., Bakhshipour, A. E., Azizipour, M., & Haghighi, A. (2025). Developing Machine Learning Models for Optimal Design of Water Distribution Networks Using Graph Theory-Based Features. Water, 17(11), 1654. https://doi.org/10.3390/w17111654

Brentan, B., Campbell, E., Goulart, T., Manzi, D., Meirelles, G., Herrera, M., & Luvizotto Jr, E. (2018). Social Network Community Detection and Hybrid Optimization for Dividing Water Supply into District Metered Areas. Journal of Water Resources Planning and Management, 144(5), 04018020. https://doi.org/10.1061/(ASCE)WR.1943-5452.0000924

Caetano, J., Carriço, N., & Covas, D. (2025). Sectorization of Water Distribution Network through a Multi-Objective Evolutionary Algorithm.

Campbell, E., Izquierdo, J., Montalvo, I., Ilaya-Ayza, A., Pérez-García, R., & Tavera, M. (2015). A Flexible Methodology to Sectorize Water Supply Networks Based on Social Network Theory Concepts and Multi-Objective Optimization. Journal of Hydroinformatics, 18(1), 62-76. https://doi.org/10.2166/hydro.2015.146

Cao, J., Wang, H., Jiao, J., Hu, K., & Qi, P. (2025). A Social Network Graph Partitioning Algorithm Based on Double Deep Q-Network. Scientific reports, 15(1), 34339. https://doi.org/10.1038/s41598-025-16768-x

Creaco, E., Giudicianni, C., & Mottahedin, A. (2023). Improved Community Detection for WDN Partitioning in the Dual Topology Based on Segments and Valves. Journal of Hydroinformatics, 25(4), 1341-1357. https://doi.org/10.2166/hydro.2023.209

Deng, Z., Lloyd, C. T., Okotto, L. G., Okotto-Okotto, J., & Wright, J. (2025). Graph and Community Detection Analysis of Pipeline Network Configuration and Urban Morphology in Kisumu and Kigali. Journal of Water Resources Planning and Management, 151(9), 05025012. https://doi.org/10.1061/JWRMD5.WRENG-6998

Di Nardo, A., Di Natale, M., Giudicianni, C., Musmarra, D., Santonastaso, G. F., & Simone, A. (2015). Water Distribution System Clustering and Partitioning Based on Social Network Algorithms. Procedia Engineering, 119(1), 196-205. https://doi.org/10.1016/j.proeng.2015.08.876

Herrera, M., Giudicianni, C., Sasidharan, M., Wright, R., Creaco, E., & Parlikad, A. K. (2025). Landmark-Node Based Reliability Assessment for Critical Infrastructure Networks. Reliability Engineering & System Safety, 111563. https://doi.org/10.1016/j.ress.2025.111563

Jojozi, F. N., Mbhele, T. P., & Eresia-Eke, C. (2025). Water Management Optimization Challenges: A Supply-Side Perspective. Sustainable Water Resources Management, 11(5), 103. https://doi.org/10.1007/s40899-025-01262-z

Ma, J., & Fan, J. (2020). Complex Network Community Detection Based on Genetic Algorithm Using K-Cliques.

Mondal, B., Bari, Q. H., & Shafiquzzaman, M. (2023). A Ternary Diagram Approach to Evaluate the Sustainability of Existing Water Supply Systems in Southwest Coastal Bangladesh. Sustainable Water Resources Management, 9(5), 143. https://doi.org/10.1007/s40899-023-00925-z

Peng, S., Cheng, J., Wu, X., Fang, X., & Wu, Q. (2022). Pressure Sensor Placement in Water Supply Network Based on Graph Neural Network Clustering Method. Water, 14(2), 150. https://doi.org/10.3390/w14020150

Peng, W., Li, Z., Li, J., & Hu, G. (2025). Reinforcement Learning Enhanced Multi-Objective Social Network Search Algorithm for Engineering Design Problems. Mathematics, 13(22), 3613. https://doi.org/10.3390/math13223613

Qi, Y., Li, Q., Cao, G., Li, G., Zhang, L., & Yang, B. (2025). Unsupervised Graph-Based Deep Clustering for Optimal Monitoring Point Placement in Urban Drainage Networks. Engineering Reports, 7(12), e70544. https://doi.org/10.1002/eng2.70544

Sodhani, S., Zhang, A., & Pineau, J. (2021). Multi-Task Reinforcement Learning with Context-Based Representations.

Spizzo, F., Venaruzzo, G., Nicolini, M., & Goi, D. (2023). Water Distribution Network Partitioning Based on Complex Network Theory: The Udine Case Study. Water, 15(8), 1621. https://doi.org/10.3390/w15081621

Sun, Q., Zhang, Y., Li, M., Liu, H., Lu, B., & Li, S. (2025). Low Model-Dependent Sensor Placement Method for Water Distribution Networks Using Adaptive Attributed Graph Clustering. Journal of Water Process Engineering, 71, 107358. https://doi.org/10.1016/j.jwpe.2025.107358

Sun, X., Peng, A., Ma, S., Jin, X., Shi, Y., Qu, Z., & Xiao, N. (2025). Multi-Regional Economic Dispatch and Demand Response Optimization in Smart Grids: A Reinforcement Learning Approach for Economic Efficiency and Strategic Decision-Making. Engineering Research Express, 7(4), 0453a0455. https://doi.org/10.1088/2631-8695/ae28bb

Talatahari, S., Bayzidi, H., & Saraee, M. (2021). Social Network Search for Global Optimization. IEEE Access, 9, 92815-92863. https://doi.org/10.1109/ACCESS.2021.3091495

Zhang, X., Zhou, X., Yu, H., Shao, Y., Ying, B., & Yu, T. (2025). Optimal Design of District Metered Areas Based on Multiple Hydraulic Performance Indicators for Water Distribution Systems. Water Supply, 25(9), 1310-1328. https://doi.org/10.2166/ws.2025.077