MITESD_77S10_lec12 (1)

MITESD_77S10_lec12 (1) - Multidisciplinary System Design...

Info iconThis preview shows pages 1–8. Sign up to view the full content.

View Full Document Right Arrow Icon
1 © Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox Engineering Systems Division and Dept. of Aeronautics and Astronautics Genetic Algorithms (cont.) Particle Swarm Optimization Tabu Search Optimization Algorithm Selection Lecture 12 Olivier de Weck Multidisciplinary System Design Optimization
Background image of page 1

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
2 © Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox Engineering Systems Division and Dept. of Aeronautics and Astronautics Today’s Topics • Genetic Algorithms (part 2) • Particle Swarm Optimization • Tabu Search • Selection of Optimization Algorithms
Background image of page 2
3 © Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox Engineering Systems Division and Dept. of Aeronautics and Astronautics Genetic Algorithms (Part 2)
Background image of page 3

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
4 © Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox Engineering Systems Division and Dept. of Aeronautics and Astronautics GA Convergence Typical Results generation global optimum (unknown) Converged too fast (mutation rate too small?) Average performance of individuals in a population is expected to increase, as good individuals are preserved and bred and less fit individuals die out. Average Fitness
Background image of page 4
5 © Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox Engineering Systems Division and Dept. of Aeronautics and Astronautics GAs versus traditional methods Differ from traditional search/optimization methods: GAs search a population of points in parallel, not only a single point GAs use probabilistic transition rules , not deterministic ones GAs work on an encoding of the design variable set rather than the variable set itself GAs do not require derivative information or other auxiliary knowledge - only the objective function and corresponding fitness levels influence search
Background image of page 5

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
6 © Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox Engineering Systems Division and Dept. of Aeronautics and Astronautics Parallel GA’s GA’s are very ameniable to parallelization. Motivations: - faster computation (parallel CPU’s) - attack larger problems - introduce structure and geographic location There are three classes of parallel GA’s: • Global GA’s • Migration GA’s • Diffusion GA’s Main differences lie in : - population structure - method of selecting individuals for reproduction
Background image of page 6
7 © Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox Engineering Systems Division and Dept. of Aeronautics and Astronautics Global GA GA Farmer Selection Assign Fitness Worker 1 Crossover Mutation Function evaluation Worker 2 Crossover Mutation Function evaluation Worker N Crossover Mutation Function evaluation ... • GA Farmer node
Background image of page 7

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Image of page 8
This is the end of the preview. Sign up to access the rest of the document.

This note was uploaded on 11/08/2011 for the course AERO 16.851 taught by Professor Ldavidmiller during the Fall '03 term at MIT.

Page1 / 39

MITESD_77S10_lec12 (1) - Multidisciplinary System Design...

This preview shows document pages 1 - 8. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online