A multi-strategy integrated improved nutcracker optimization algorithm and its applications in water conservancy engineering

To enhance the performance of nutcracker optimization algorithm (NOA) for complex optimization problems in water conservancy engineering, an improvement framework is investigated. The performances of the selected 26 meta-heuristic algorithms are analyzed using benchmark functions. Targeting the limitations of NOA, the improved NOA (INOA) is constructed by incorporating logistic-tent chaotic mapping, tangent flight strategy, dynamic fitness-distance balance selection, criss-cross mutation, Gaussian mutation perturbation, and Tent chaotic perturbation. INOA is adopted to optimize attention mechanism long short-term memory neural network (ATLSTM) for broken wire signal identification of prestressed concrete cylinder pipe (PCCP) and to optimize the hydraulic-seasonal-time (HST) model for dam deformation prediction. The results indicate that the NOA exhibits optimal overall performance among the selected algorithms and that the INOA improves optimization accuracy, convergence speed, stability, and operational efficiency. In both engineering applications, INOA outperforms other baseline optimizers, demonstrating its potential for intelligent monitoring and prediction in water conservancy engineering.