1 / 15

Derivative Free Optimization

Derivative Free Optimization. G.Anuradha. Contents. Genetic Algorithm Simulated Annealing Random search method Downhill simplex method. Characteristics of Derivative free optimization techniques. Derivative freeness Does not require derivatives

benjamin
Download Presentation

Derivative Free Optimization

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Derivative Free Optimization G.Anuradha

  2. Contents • Genetic Algorithm • Simulated Annealing • Random search method • Downhill simplex method

  3. Characteristics of Derivative free optimization techniques • Derivative freeness • Does not require derivatives • Rely on repeated evaluations of objective function • Intuitive guidelines • Guidelines based on simple intuition • Some concepts are based on nature’s wisdom and thermodynamics • Slowness • Comparatively slower to derivative based optimization techniques.

  4. Characteristics of Derivative free optimization techniques Contd… • Flexibility:- • Works well with a complex objective function also. • Randomness • Derivative free methods are stochastic in nature • They are global optimizers given enough computational time • Analytic Opacity • Difficult to do analytic studies of these methods • Most of the knowledge about these methods are empirical in nature.

  5. Characteristics of Derivative free optimization techniques Contd… • Iterative Nature:- • Unlike LSE these methods are iterative in nature and requires certain stopping criteria to determine when to terminate the optimization process • For eg. The stopping criteria for a maximization problem includes • Computational time • Optimization goal • Minimal improvement • Minimal relative improvement

  6. Genetic Algorithm • Genetic Algorithms are search and optimization techniques based on Darwin’s Principle of Natural Selection. • Proposed by John Holland at University of Michigan in 1975

  7. Darwin’s Principle Of Natural Selection IF there are organisms that reproduce, and IF offsprings inherit traits from their parents, and IF there is variability of traits, and IF the environment cannot support all members of a growing population, THEN those members of the population with less-adaptive traits (determined by the environment) will die out, and THEN those members with more-adaptive traits (determined by the environment) will thrive The result is the evolution of species.

  8. Basic Idea Of Principle Of Natural Selection “Select The Best, Discard The Rest”

  9. Principles behind GA • Encodes each point in solution space into binary bit string called a chromosome • Each point is associated with a fitness value • GA keeps a set of points called population or gene pool. • In each generation GA constructs a new population using crossover and mutation • Members of higher fitness values are most likely to survive and to participate in mating or crossover operations. • This is analogous to the Darwin’s model of evolution

  10. How is it different from other optimization and search procedures? • Works with a coding of the parameter set, not the parameters themselves • Search for a population of point and not a single point • Use objective function information and not derivatives or other auxiliary knowledge • Uses probabilistic transition rules and not deterministic rules

  11. Characteristics of GA

  12. Major components of GA • Encoding schemes • Fitness evaluation • Parent selection • Crossover operators • Mutation operators

  13. Crossover operators

  14. Elitism A policy of always keeping a certain number of best members when each new population is generated is called ELITISM

More Related