Peregrine falcon predation algorithm: a better solution to multiple engineering problems and lithium-ion battery model parameter identification problem

LL Chen and TB Liu and PF Zhao and BH Yuan and YL Hu, CLUSTER COMPUTING- THE JOURNAL OF NETWORKS SOFTWARE TOOLS AND APPLICATIONS, 28, 998 (2025).

DOI: 10.1007/s10586-025-05682-6

Heuristic algorithms have become a research hotspot due to their applicability and flexibility. This paper proposes a swarm-based optimization algorithm named Peregrine Falcon Predation Algorithm (PFPA). The PFPA is inspired by simulating the high-speed hunting behaviors of peregrine falcon couple, such as hover searching, cooperating tracking and prey capture, and fledgling training. The effectiveness of PFPA is evaluated by comparing with 10 other algorithms using the 43 benchmark functions from CEC2005, CEC2019 and CEC2020. Experimental results show that the PFPA outperforms all other algorithms on 87% of the functions from CEC2005, 30% from CEC2019, and 70% from CEC2020. The applicability of PFPA has been is further tested on five structural optimization problems and the lithium-ion battery parameter identification problem. The results demonstrate that PFPA exhibits superior global search capability and stability than the comparison algorithm. Notably, PFPA significantly enhances the accuracy of lithium- ion batteries parameter identification under extreme state of charge (SOC) levels. Specifically, during the period which exceeds the voltage plateau, PFPA improves identification accuracy by 71.58%, 22.26%, 8.03%, 10.91%, 22.26%, 27.20%, 48.88%, 11.65%, 27.16%, and 30.65% compared to GSA, TSA, SCA, HOA, NGO, PSO, FA, WOA, TOC, and GA algorithms, respectively. During the period below the voltage plateau, the identification accuracy improves by 20.04%, 7.28%, 4.87%, 7.69%, 7.39%, 8.80%, 7.63%, 7.74%, 7.39%, and 20.57%, respectively. These results confirm that the PFPA possesses outstanding search capability, strong ability to escape from local optima, and significant competitive advantages in solving real-world optimization problems with multiple constraints and variables.

Return to Publications page