Building Protection Model Based on Puma Optimizer and LSTM Neural Network

We established the Comprehensive value evaluation system for ancient architecture by reviewing the information, and the system has three main indicators: cultural and historical value, the intrinsic value of We take Gansu ancient architecture as the basis and determine the weights by TOPSIS-entropy weighting method, and the weights of the indexes are 0.4970, 0.2539, 0.2491, respectively. After that, we choose Jiayuguan Pass, Mogao Grottoes, and Zhuge Military Bastion as the objects of the study, and we use the appropriation of maintenance to protect the ancient architecture, and use Puma optimization to protect the ancient architecture. The Puma optimizer is used to determine and predict the appropriations for these three ancient buildings from 2024 to 2034 using LSTM. It is proposed to give the most allocation to Dunhuang Mogao Caves area, which will be allocated 1.21473 billion yuan in 2034, and to give a smaller allocation to Zhuge Barracks area, which will be allocated 0.0203 billion yuan in 2034.