为提升混联水库群联合防洪能力、降低洪灾损失,协同控制水库最高运行水位与下游控制断面最大过流量,以“防洪系统综合安全度最大”为准则构建混联水库群防洪优化调度模型;提出平滑解策略,耦合差分进化思想、Levy突变策略,构建改进哈里斯鹰优化算法,通过经典多极值测试函数验证其在寻优精度与收敛稳定性方面的优越性。以湘江上游流域的涔天河、双牌、欧阳海、东江四座混联水库群为研究对象,开展联合防洪优化调度,将改进哈里斯鹰算法与传统哈里斯鹰算法优化结果进行对比。结果表明,改进哈里斯鹰优化算法在削峰率、错峰时效和水库水位控制等方面均表现更优,验证了其在水库群防洪优化调度中的优越性与工程适用性。

[Objective] This study addresses the complex challenge of optimizing flood control operations for hybrid-connected reservoir systems. It aims to develop a coordinated strategy for managing both reservoir water levels and downstream discharges to enhance joint flood control capacity and reduce flood losses. A core objective is to create an advanced optimization algorithm capable of solving the high-dimensional and constrained model efficiently. The research is ultimately validated through a real-world case study to demonstrate practical superiority over existing methods. [Methods] The research methodology integrates model formulation with algorithmic innovation. First, an optimization model is established with the objective of maximizing the comprehensive safety degree of the flood control system. To solve this complex model, a novel Improved Harris Hawks Optimization (IHHO) algorithm is proposed. The improvements include a smooth solution strategy that pre-processes discharge sequences to eliminate impractical fluctuations and guide the search; the integration of differential evolution concepts during the global exploration phase to enhance population diversity and guidance; and a dynamic Lévy mutation strategy during local development to refine search precision. The algorithm’s superior performance in optimization accuracy and convergence stability is first verified using standard multi-extremum test functions. Finally, the model and IHHO are applied to a case study of four major reservoirs (Centianhe, Shuangpai, Ouyanghai, and Dongjiang) in the upper Xiangjiang River basin, with results compared against the standard HHO algorithm. [Results] The application of IHHO to the Xiangjiang River case study yielded significantly better operational outcomes compared to the standard HHO. In terms of flood peak attenuation, the IHHO-optimized scheme achieved notably higher peak reduction rates at the critical downstream control sections. Specifically, the peak shaving rate at the Hengyang section reached 21.41%, substantially exceeding the 14.38% achieved by HHO. Effective peak timing delay was also accomplished. Regarding reservoir operation safety, IHHO successfully implemented a coordinated water level control strategy aligned with predefined priority weights. It maintained lower maximum levels for key high-weight reservoirs like Dongjiang and Centianhe, while reasonably utilizing the capacity of lower-weight reservoirs like Ouyanghai within safe limits to gain greater system-wide benefits. Furthermore, the smooth solution strategy inherent to IHHO ensured that the optimized discharge processes were stable and practically feasible, free from the abrupt fluctuations observed in the HHO results, highlighting better convergence stability and solution quality. [Conclusions] This study demonstrates the effectiveness of the proposed Improved Harris Hawks Optimization algorithm for the optimal flood control operation of hybrid-connected reservoir groups. The algorithmic enhancements—smooth solution processing, differential evolution, and dynamic Lévy mutation—collectively confer superior global search capability, local exploitation precision, and convergence stability for solving such high-dimensional, constrained problems. When applied with the “comprehensive safety degree” model, IHHO generates operation schemes that outperform traditional methods across key performance metrics: downstream flood peak reduction, peak delay, and coordinated, safe reservoir level regulation. The case study confirms the practical applicability and superiority of this integrated approach.