دانشگاه آزاد اسلامی واحد تهران جنوب

صفحه اصلی / نخستین همایش هوش مصنوعی و پژوهش های نوظهور: همگرایی انسان و سیستم های هوشمند

A CRISP-DM-Based Multi-objective Optimization of Building Wall Panels Using PSO and SA

نویسندگان :

Mohammad Chehrzad¹ Seyed Azim Hosseini² Razieh Farazkish³ Mohammad Emami korandeh⁴

1- دانشجوی دانشگاه آزاد اسلامی واحد تهران جنوب 2- دانشگاه آزاد تهران جنوب 3- دانشگاه آزاد تهران جنوب 4- دانشگاه آزاد تهران جنوب

کلمات کلیدی :

Building Energy Efficiency،CRISP-DM methodology،Particle Swarm Optimization (PSO،Simulated Annealing (SA)

چکیده :

The building sector accounts for a substantial share of global energy consumption, with a significant portion lost through external walls. This study, based on the CRISP-DM framework, proposes a multi-objective optimization model using Particle Swarm Optimization (PSO) and Simulated Annealing (SA) to enhance wall design. Three criteria,thermal resistance, total thickness, and weight,were optimized using a refined dataset of 207 materials. Materials were pre-classified, and penalties were applied to ensure construction feasibility and material diversity. Among 120 panel combinations, PSO showed better accuracy and convergence (mean thermal deviation: +0.13 m²·K/W), while SA favored lighter solutions with slightly higher thermal resistance (mean: +0.18 m²·K/W). PSO explored a broader material range, whereas SA produced more focused selections. PSO is thus recommended for accuracy-driven applications, with SA as a complementary tool for exploring innovative solutions