Hamid Reza Vosoughifar - Asst. Professor, Faculty of Engineering, Islamic Azad University, South Tehran Branch
Arezoo Mohammadi - Student of Master of Science, Department of civil engineering, University of Zanjan
Seyed Kazem Sadat Shokouhi

In recent years, light weight steel framing system has been proposed as an economic system and earthquake-resistant. Tendency of the mass constructors to this system is due to being full industrialprocess. One of the resistant systems against lateral load in cold-formed steel structure is applying of braces which optimization and connections improvement for these braces have been considered by experts of this field research. In this paper, different experimental studies and normalization andsimulation by ANN were used. The results of this research have been applied for create a nonlinearrelationship. First all of data such as input, target must be normalized and then simulating and training by neural network should be done. In this research, two layers have been used. One of these layers is sigmoidlayer. Results show that optimal connections in light weight steel framing system have suitable plasticity, load capacity and nonlinear relation. Statistical analysis results on SPSS software show that there is no significant different between neural network and experimental results (P-Value > 0.05

Cold-formed steel (CFS), Light weight steel frame (LSF), Optimization, Neural network