Cooperatively coevolving simulated annealing for optimization design of the concentrated-force diffusion component

Abstract

Load-carrying capacity and load-diffusion performance are two primary factors to design the concentrated-force diffusion component (CFDC) in launch vehicles. To improve both of the two factors simultaneously, a concept of the hierarchical stiffened shell, along with a novel structural configuration, is first utilized to tailor the CFDC subjected to axial concentrated load. Since the explosion of design variables, the burden of post-buckling optimization will be very prodigious and unaffordable while designing such hierarchical stiffened shells. To mitigate the computational burden, a novel analysis method incorporating the linear static method and Euler buckling strength estimation is proposed for the construction of the optimization problem of CFDC. Moreover, by introducing the concept of cooperative coevolution to the simulated annealing algorithm, a novel cooperatively coevolving simulated annealing (CCSA) algorithm is proposed for the facilitation of the optimization, in which two types of structural decomposition strategies are proposed for decomposing the hyper multi-dimensional problem into several lower-dimensional subcomponents. The performance benefits of the proposed methods for enhancing optimization gains and efficiency are validated by the competitive optimal results of the CFDC with the traditional SA.