SUMMARYOne variant of the ant colony optimization (ACO) metaheuristic, known as rank-based ant system (RBAS), is proposed for weight minimization of structures involving discrete design variables. Stress and displacements constraints are handled using a penalty technique, and the structural analysis is performed by the finite element method.Results obtained for several discrete weight minimization problems show that the RBAS algorithm implemented is effective and competitive when compared with results found by genetic algorithms and another RBAS variant.

Clonal selection algorithms are computational methods inspired by the behavior of the immune system which can be applied to solve optimization problems. However, like other nature inspired algorithms, they can require a large number of objective function evaluations in order to reach a satisfactory solution. When those evaluations involve a computationally expensive simulation model their cost becomes prohibitive. In this paper we analyze the use of surrogate models in order to enhance the performance of a clonal selection algorithm. Computational experiments are conducted to assess the performance of the presented techniques using a benchmark with 22 test-problems under a fixed budget of objective function evaluations. The comparisons show that for most cases the use of surrogate models improve significantly the performance of the baseline clonal selection algorithm.

No abstract

No abstract

No abstract