web ga webzueri

Transcript

1. Genetic Algorithm for Web Developers Jean-Baptiste Clion Google Developer Expert Web Zürich
2. 10 01 11 00 01 10 00 01 11 10 00 01 11 10 00 00
3. 01100011001001000100110100010101 32 bits = 232 = 4,294,967,296 possibilities four billion, two hundred and ninety-four million, nine hundred and sixty-seven thousand, two hundred and ninety-six
4. Search space The set of all possible solutions able to satisfy a given problem constraints.
5. 00000000000000000000000000000000 TO 11111111111111111111111111111111
6. Find a single value
7. Intelligence Capacity to search the most adapted solution to a given problem in the most efficient way and within the shortest possible time.
8. 01010010110000010000110010010000 | 0.24 11001001100011100000111100000011 | 0.48 01010010110111010010110010010000 | 0.64 Fitness operator
9. 00000000000000000000000000000000 11111111111111111111111111111111 Crossover operator Solution #1 Solution #2 00110011001100110011001100110011 Solution #3 Merge two solutions based on a given pattern. [ ]
10. 00110011001100110011001100110011 Mutate operator 00100011011100110011001100100011 [Flip bit] [Flip bit] Inverse bits on a given probability [ ] Mutation ratio Solution #1 Solution #2
11. 1. Select N random possible solutions in space 2. Appy Fitness function on selected solutions 3. Sort selected solution by Fitness 4. Extract best solution 5. Is best a valid solution? Yes => STOP 6. Crossover best solution with other solutions 7. Mutate new solutions 8. Add best to avoid regression
12. Twitter @JBaptisteClion Evoria - GitLab https://gitlab.com/JBClion/Evoria Speakerdeck https://speakerdeck.com/jbclion/web-ga-webzueri Evoria https://evoria-1984.firebaseapp.com All you need! Example - GitLab https://gitlab.com/JBClion/GA-SIMPLE Long Talk https://www.youtube.com/watch?v=_xVlLHHb-Nw