Evaluation Model for Public Building Design Based on Artificial Intelligence Algorithms

Due to the complexity of comprehensive evaluation in public building design, as well as the nonlinear relationship among influencing factors, traditional evaluation methods face two key challenges: reducing evaluation errors and enhancing the effectiveness and rationality of evaluation models. In this study, a comprehensive evaluation model for multi-attribute schemes is proposed using modern optimization algorithms. The model incorporates optimized general methods for structuring tables, scheme evaluation tables, and standardized calculation models, forming an evaluation index system for public building design schemes. Additionally, improved supervised input-based BP neural networks and self-organizing competitive neural network evaluation models and algorithms, utilizing unsupervised input, are introduced. Results indicate that both neural network evaluation models demonstrate algorithmic effectiveness and rationality, making them suitable for the comprehensive evaluation of public building design schemes. Moreover, these models address the main issues encountered in previous evaluation methods and reduce subjectivity in each evaluation stage. The designed neural network evaluation models and algorithms exhibit strong generalization abilities, offering convenience, accuracy, reliability, and wide-ranging application prospects. They represent significant advancements and supplements to existing theories in the field of architectural design evaluation and hold reference value for evaluation work in other domains.