Abstract
Five algorithms for the simultaneous optimal design of smart structural systems are presented. These algorithms are developed based on sequential mathematical programming and guided random search techniques, being applied to multidisciplinary optimization in which both structural layout and controller parameters are involved. A method to prevent singularities when updating the structural layout is suggested. Two adaptive trusses are optimised and comparison is made for the different optimal schemes on the basis of iteration histories and major results achieved. All of the five algorithms are shown efficient in improving the truss's performance with respect to robustness and controllability. The results show that the greatest improvement in all the performance indicators is achieved using a genetic algorithm, whilst the most efficient scheme is seen to be a novel hybrid combining sequential linear programming with simulated annealing.
Original language | English |
---|---|
Pages (from-to) | 25-37 |
Number of pages | 13 |
Journal | Computer Methods in Applied Mechanics and Engineering |
Volume | 184 |
Issue number | 1 |
DOIs | |
Publication status | Published - 31 Mar 2000 |
Keywords
- Genetic algorithm
- Multi-objective programming
- Sequential mathematical programming
- Simulated annealing
- Simultaneous structural/control system optimization
- Smart structure