Amirkabir University of TechnologyAmirkabir Journal of Civil Engineering2588-297X53520210723Sizing and Geometry Optimization of Truss Structures Using a Hybrid of Gravitational Search Algorithm and Cellular AutomataSizing and Geometry Optimization of Truss Structures Using a Hybrid of Gravitational Search Algorithm and Cellular Automata21492174417510.22060/ceej.2020.17362.6538FAMiladDehghaniDepartment of Civil Engineering, Faculty of Engineering, University of Qom,Qom, IranMostafaMashayekhiDepartment of Civil Engineering, Faculty of Technical and Engineering, Vali-e-Asr University of Rafsanjan, Rafsanjan, Iran0000-0002-0859-5593MehdiSharifiAssistant Prof, Department of Civil Engineering, Faculty of Technical and Engineering, University of Qom, Qom, Iran0000-0001-8166-1668Journal Article20191111<span style="letter-spacing: .05pt;">In this study, a new method is presented to solve the geometry and sizing optimization problems of truss structures using an effective </span>hybrid <span style="letter-spacing: .05pt;">of cellular automata (CA) and gravitational search algorithm (GSA), which is named the CA-GSA method. The basic of the GSA is the Newtonian Gravity and Motion laws. Due to the direct effect of all objects on each other and the lack of attention to elitist selection, this algorithm converges to a local optimum point. In this study, with the help of the CA method, masses are distributed in a finite cellular network, and each cell is only related to its neighbors. In the CA-GSA method, the laws of gravity and motion of masses in the GSA method are defined as the relationship factor of each cell to its neighboring ones. Therefore, the applied force on each mass is obtained from the resultant force of its top neighboring masses. The definition of these top neighboring masses and their applied force on the central mass add memory and elitist selection to the GSA algorithm, respectively. Another advantage of the new method is to update the cellular network after any local evolution, which makes it possible to achieve the optimal point using fewer analyzes. To investigate the usefulness of the proposed method, the CA-GSA method was used to solve the geometry and sizing optimization problems of four benchmark truss structures. The results of CA-GSA show the superiority and power of this algorithm in comparison with the methods introduced in the literature.</span><span style="letter-spacing: .05pt;">In this study, a new method is presented to solve the geometry and sizing optimization problems of truss structures using an effective </span>hybrid <span style="letter-spacing: .05pt;">of cellular automata (CA) and gravitational search algorithm (GSA), which is named the CA-GSA method. The basic of the GSA is the Newtonian Gravity and Motion laws. Due to the direct effect of all objects on each other and the lack of attention to elitist selection, this algorithm converges to a local optimum point. In this study, with the help of the CA method, masses are distributed in a finite cellular network, and each cell is only related to its neighbors. In the CA-GSA method, the laws of gravity and motion of masses in the GSA method are defined as the relationship factor of each cell to its neighboring ones. Therefore, the applied force on each mass is obtained from the resultant force of its top neighboring masses. The definition of these top neighboring masses and their applied force on the central mass add memory and elitist selection to the GSA algorithm, respectively. Another advantage of the new method is to update the cellular network after any local evolution, which makes it possible to achieve the optimal point using fewer analyzes. To investigate the usefulness of the proposed method, the CA-GSA method was used to solve the geometry and sizing optimization problems of four benchmark truss structures. The results of CA-GSA show the superiority and power of this algorithm in comparison with the methods introduced in the literature.</span>https://ceej.aut.ac.ir/article_4175_e0438ae81e41a02b890588238dead793.pdf