Skip to content

Latest commit

 

History

History
13 lines (9 loc) · 818 Bytes

README.md

File metadata and controls

13 lines (9 loc) · 818 Bytes

global-optimization

Unknown function approximation, given input-output measurements, using Genetic Algorithm

Preliminary implementation of a Genetic Algorithm, utilized to approximate an unknown function, given input-output measurements. Flexibility is provided to obtain different results and bias the approximation towards a desired direction, through selection of multiple parameters such as, the size of the population, the k more accurate chromosomes to pick, the number of generations, and the tester-set of the input values.

note

For optimal approximation, the threshold (desired MSE) must be kept low enough, anywhere in the range (10^-5, 10^-7). This however, translates to excessive computational requirements, and a pretty poor performance.