The evaluation of application effects of multi-algorithm collaborative architecture in distributed hydrological models

In response to the increasing complexity of hydrological modeling demands driven by climate change and human activities, this study introduces a Multi-Algorithm Collaborative Architecture (MACA) to enhance parameter calibration in distributed hydrological models, specifically utilizing the HBV-92 model in the upper Shaying River basin, Huai River Basin, China. MACA integrates multiple optimization algorithms, including Genetic Algorithm (GA), Differential Evolution (DE), Particle Swarm Optimization (PSO), Shuffled Complex Evolution (SCE-UA), and Chaotic Particle Swarm Genetic Algorithm (CPSGA), to comprehensively explore the parameter space and overcome the limitations of single optimization methods. Systematic evaluation against five single optimization algorithms demonstrates that MACA significantly outperforms them in simulation accuracy, stability, computational time, and convergence characteristics. It achieves a Nash Efficiency Coefficient (NSE) of 0.83, a 13.70% improvement over the baseline GA, indicating significantly higher precision in capturing flow variations and peak flows. In terms of stability, MACA shows superior performance with the lowest coefficient of variation at 200 iterations, indicating high repeatability and consistency. Despite its structural complexity, MACA maintains remarkable computational efficiency, with a runtime comparable to simpler algorithms like GA and DE, demonstrating no significant time penalty for its superior performance. These findings demonstrate MACA’s potential as a robust tool for simulating complex hydrological processes, providing scientific insights for water resource management and flood disaster prevention.