China Safety Science Journal ›› 2025, Vol. 35 ›› Issue (S1): 246-251.doi: 10.16265/j.cnki.issn1003-3033.2025.S1.0037

• Original article • Previous Articles     Next Articles

Demand prediction of earthquake disaster rescue equipment based on ISSA-BP

LIU Hao1(), SHI Fuli2, LUO Lei2, LI Wenbo3, LI Wenyuan4   

  1. 1 Graduate Regiment, Engineering University of People's Armed Police, Xi'an Shaanxi 710086, China
    2 College of Equipment Management and Support, Engineering University of People's Armed Police, Xi'an Shaanxi 710086, China
    3 Gansu Earthquake Agency, Lanzhou Gansu 730000, China
    4 Police Dog Technical College, Criminal Investigation Police University of China, Shenyang Liaoning 110854, China
  • Received:2025-01-14 Revised:2025-03-17 Online:2025-06-30 Published:2025-12-30

Abstract:

To enhance the efficiency of earthquake rescue equipment allocation and support, historical earthquake rescue information in China was analyzed, with the number of affected people as the prediction object and eight disaster information, such as magnitude, focal depth, and seismic intensity, as influencing factors. An ISSA based on BP neural network, spatial pyramid matching (SPM) chaotic mapping, sine-cosine algorithm, and Levy flight strategy was proposed. Combined with the quantitative relationship between the number of affected people and rescue equipment, the demand for earthquake rescue equipment was indirectly predicted. The ″12·18 Jishishan Earthquake″ rescue case was used for verification. The results show that the ISSA-BP model has higher accuracy in predicting the number of affected people and can effectively predict the number of affected people after an earthquake, thereby estimating the required rescue equipment. The ″12·18 Jishishan Earthquake″ rescue case verifies the practicality of the model in predicting the demand for rescue equipment after an earthquake.

Key words: improved sparrow search algorithm (ISSA), back propagation (BP), earthquake disaster, rescue equipment, demand prediction

CLC Number: