A novel global optimization method of truss topology

This paper proposes a global topology optimization algorithm based on subset simulation for the singular optimum problem subject to stress constraints of trusses. The constraints are handled by a fitness function which reflects their degree of violation. The rational and global topology results are guaranteed by the judgment of the samples’ rationality and the Metropolis-Hasting algorithm. Three examples show that the established method can quickly reduce the searching region to the feasible region and converge to the global optimum precisely enough for the singular optimum problem.