Stochastic directional search: An ecient heuristic for structural optimization of building frames

Performance of an optimization heuristic highly depends on the method it employs to represent and decompose the search space. A class of optimization methods, including swarm intelligence, utilizes a special way of decomposing the search space via specied direction states. The present work expands the idea, providing a stochastic directional state search with tuned thresholds for selecting every direction state. The proposed meta-heuristic is then utilized for the problem of structural weight minimization in building frames under lateral and gravitational loading using a nite element analyzer. Treating a number of examples from the literature, the eciency and eectiveness of the proposed method are shown to be superior to some other meta-heuristics, including an improved particle swarm optimizer.