A New Flood Routing Framework Based on Modified Muskingum Model and Nature-Based Optimization Algorithms

Abstract

This study presents a new flood routing method integrating the modified Muskingum (NLM7_Aqlat) method with hybrid natural optimization algorithms (hybrid of Humboldt squid optimization algorithm [HSOA] and gradient-based optimizer [GBO] and hybrid of Pine cone optimization algorithm [PCOA] and GBO). In the NLM7_Aqlat, the lateral flow is applied to a seven-parameter nonlinear Muskingum model (NLM7), and hybrid natural-based optimization algorithms optimize the parameters. In Karahan flood routing, the standard value of the mean sum of squared deviations (SSQ_mean) for integrating the NLM7_Aqlat model and PCOA_GBO was calculated to be 96.06% less than the other 10 algorithms (such as GA and GBO). In Wilson flood routing, the PCOA_GBO algorithm in the NLM7 model calculated the SSQ_mean criterion value 99% lower than other optimization algorithms. The HSOA_GBO algorithm in the NLM7_Aqlat model provided the best flood routing for Weisman-Lewis, enhancing hydrograph accuracy. In Karun flood routing, the PCOA algorithm estimated the SSQ_mean in the NLM7 model to be 89% lower than other algorithms. The new flood routing method showed competitive results versus NLM7. Hybrid optimization algorithms outperformed standalone ones, prompting authors to recommend this methodology for enhancing early flood warning systems.