Semnan University Press Journal of Modeling in Engineering 2008-4854 17 57 2019 06 22 A fast hybrid refining optimization with sensitivity-based partitioning approach for truss structures A fast hybrid refining optimization with sensitivity-based partitioning approach for truss structures 127 146 3914 10.22075/jme.2019.14706.1457 FA Saeed Farahi Shahri PhD Student, Civil Engineering Department, University of Sistan and Baluchestan, Zahedan, Iran Mohammad Reza Ghasemi Professor, Civil Engineering Department, University of Sistan and Baluchestan, Zahedan, Iran Seyed Roohollah Mousavi Associate Professor, Civil Engineering Department, University of Sistan and Baluchestan, Zahedan, Iran Journal Article 2018 05 08 So far, several methods have been utilized for truss optimization where metaheuristic thechniqes have been used as one of the most common methods for such purpose. These thechniqes have received much attention due to their independence from gradient information and efficient global search property. In some of the metaheuristic thechniqes, random generation of the preliminary population leads to creation of repetitive population, dependency of final optimum solution on preliminary population, increasing of the number of analyses and reduction of the algorithm efficiency. To overcome these disadvantages, the present study employs some intelligent thechniqes for the generation and the selection of the population. So, a new metaheuristic algorithm called SPR is proposed for discrete truss optimization problems. The SPR algorithm is based on the sensitivity of the elements, the partitioning of the search space and the rebirthing of the optimization procedure. The structural members which are more sensitive to changing the cross-sectional areas are recognized through the procedure. In order to find the best approximate solution for any design variable, the search space is discritized to some partitions. Rebirthing technique searches for better optimum solutions around the last local optimum point for any design variable to end the stagnation states. The performance of the proposed algorithm is investigated using several benchmark size optimization problems of truss structures from the literature. In comparison to the available results in the literature, the proposed SPR algorithm showed capable of obtaining the optimum design with much lesser computational attempt. So far, several methods have been utilized for truss optimization where metaheuristic thechniqes have been used as one of the most common methods for such purpose. These thechniqes have received much attention due to their independence from gradient information and efficient global search property. In some of the metaheuristic thechniqes, random generation of the preliminary population leads to creation of repetitive population, dependency of final optimum solution on preliminary population, increasing of the number of analyses and reduction of the algorithm efficiency. To overcome these disadvantages, the present study employs some intelligent thechniqes for the generation and the selection of the population. So, a new metaheuristic algorithm called SPR is proposed for discrete truss optimization problems. The SPR algorithm is based on the sensitivity of the elements, the partitioning of the search space and the rebirthing of the optimization procedure. The structural members which are more sensitive to changing the cross-sectional areas are recognized through the procedure. In order to find the best approximate solution for any design variable, the search space is discritized to some partitions. Rebirthing technique searches for better optimum solutions around the last local optimum point for any design variable to end the stagnation states. The performance of the proposed algorithm is investigated using several benchmark size optimization problems of truss structures from the literature. In comparison to the available results in the literature, the proposed SPR algorithm showed capable of obtaining the optimum design with much lesser computational attempt. https://modelling.semnan.ac.ir/article_3914_cc4baaa594a9004261a2bc238af6218b.pdf