Enhanced crow search algorithm for optimum design of structures

In this study, the capability of recently introduced crow search algorithm (CSA) was evaluated for structural optimization problems. It is observed that the standard CSA was led to undesirable performance for solving structural optimization problems. Accordingly, three modifications were made on the standard CSA to obtain the enhanced crow search algorithm (ECSA) while no parameter was added to the ECSA. First, each violated decision variable was replaced by corresponding decision variable of the global best solution. Second, a free-fly mechanism was suggested for constraint handling. Third, the personal upper bound strategy (PUBS) was proposed for elimination of inessential structural analyses. To assess the efficacy of the proposed modifications, four popular benchmark structures were employed and each modification was added to the CSA in a separate stage and then its effects were illustrated. The results of benchmark structures were examined in terms of minimum weight, convergence rate, and reliability. The results confirmed that the ECSA was significantly better than the standard CSA. Moreover, the ECSA obtained better or very competitive results in comparison with well-known and other newly developed metaheuristic methods. (C) 2019 Elsevier B.V. All rights reserved.