Reliability-based design optimization of nonlinear inelastic trusses using improved differential evolution algorithm

A robust method for sizing reliability-based design optimization (RBDO) of truss structures is developed by integrating nonlinear inelastic analysis, a structural reliability analysis method, and a proposed optimization method based on differential evolution (DE) algorithm. Both discrete and continuous optimizations are considered. Compared to conventional DE, three main improvements are introduced in the proposed optimization method such as: (1) a new mutation scheme based on p-best strategy to balance between local and global searches; (2) modifications of the scale factor and crossover rate parameters; (3) a multi-comparison technique to reduce the useless objective function evaluations. 20 test functions and 4 real-world examples are first considered to demonstrate the performance of the proposed optimization method for deterministic optimization problems. 3 nonlinear inelastic truss structures are then carried out to investigate the robustness of proposed RBDO method.