Abstract
A multilevel hybrid genetic algorithm, which integrates genetic algorithms and traditional gradient-based optimal design methods, is investigated for solving the hierarchical optimization problem in structural control system design. A three-level mathematical model for the problem is developed in which the first level optimizes the controller number, the second level optimizes the positions of controllers in a structure, and the third level optimizes the parameters of controllers. The algorithm is applied to a 16-storey building excited by typical earthquake forces and controlled by a type of passive control device.
