Scholarlink

<p>Efficient management of water reservoirs is crucial for ensuring both water quantity and quality under increasing environmental and socio-economic pressures. This study addresses the dual challenge of quantitative–qualitative reservoir operation by integrating simulation and optimization techniques. The CE-QUAL-W2 model was employed to simulate hydrodynamic and water quality behavior, with a focus on Total Dissolved Solids (TDS) concentration. To reduce computational time, a Supervised Learning (SL) surrogate model based on an artificial neural network was developed and coupled with the Archimedes Optimization Algorithm (AOA), forming a hybrid AOA–SL framework. The Ekbatan Reservoir in western Iran was selected as a case study. Results show that integrating the SL surrogate reduced the number of objective function evaluations by 25% (from 4000 to 3000) and halved the optimization run-time. The hybrid model achieved a prediction accuracy of <i>R</i><sup>2</sup> = 0.87 compared with CE-QUAL-W2 outputs. Sensitivity analysis of objective weights indicated that the configuration w₁ = 0.6 (TDS minimization) and w₂ = 0.4 (demand satisfaction) achieved the most balanced trade-off between water quality and supply reliability. Compared with GA–SL and PSO–SL models, the proposed AOA–SL framework demonstrated superior convergence stability and computational efficiency. These findings suggest that the hybrid surrogate-assisted optimization approach provides a practical and scalable solution for sustainable reservoir operation. Future applications may extend the model to multi-reservoir systems and include additional objectives such as ecological flow and energy generation.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Development of the AOA-SL surrogate model for optimal reservoir operation considering quantitative–qualitative objectives

  • 期刊论文

  • 发表日期：

    2025年12月30日

查看全文
  • Seyedeh Zahra Hassani,
  • Parisa-Sadat Ashofteh,
  • Babak Zolghadr-Asli

摘要

Efficient management of water reservoirs is crucial for ensuring both water quantity and quality under increasing environmental and socio-economic pressures. This study addresses the dual challenge of quantitative–qualitative reservoir operation by integrating simulation and optimization techniques. The CE-QUAL-W2 model was employed to simulate hydrodynamic and water quality behavior, with a focus on Total Dissolved Solids (TDS) concentration. To reduce computational time, a Supervised Learning (SL) surrogate model based on an artificial neural network was developed and coupled with the Archimedes Optimization Algorithm (AOA), forming a hybrid AOA–SL framework. The Ekbatan Reservoir in western Iran was selected as a case study. Results show that integrating the SL surrogate reduced the number of objective function evaluations by 25% (from 4000 to 3000) and halved the optimization run-time. The hybrid model achieved a prediction accuracy of R2 = 0.87 compared with CE-QUAL-W2 outputs. Sensitivity analysis of objective weights indicated that the configuration w₁ = 0.6 (TDS minimization) and w₂ = 0.4 (demand satisfaction) achieved the most balanced trade-off between water quality and supply reliability. Compared with GA–SL and PSO–SL models, the proposed AOA–SL framework demonstrated superior convergence stability and computational efficiency. These findings suggest that the hybrid surrogate-assisted optimization approach provides a practical and scalable solution for sustainable reservoir operation. Future applications may extend the model to multi-reservoir systems and include additional objectives such as ecological flow and energy generation.