Tit for Tat
Article written by: Emanuele Cassani	Home pubblicazioni

Evolution Strategies - release 2.0

Release 2.0 of the program Evolution Strategies, configurable script to perform simulations
written by Emanuele Cassani
to run the JS program it is necessary to enable the script on your browser.

The 2.0 version, new feature

to define the strategies, the ordinary aminoacids genetic code has been used
inserted new genes for: traitor Y/N, regret Y/N, how may times it will betray?
inserted new genes for: remember prefious rounds? Y/N how many do you remember?
modified the DNA with STOP genes
The mutant uses the new genes to define the strategy
the mutuations are random on any character
the name of the mutant indicates the number of changes
inserted a graphical representation for the strategies in the single turnament
inserted a graphical representation for the DNA in the single turnament
the games results are displayed in a table

Version 2.0 draft script

Single turnament
run the version 2.0 of the script

Generations turnament generazioni
run the version 2.0 of the script

Release 2.1 still to do

For the previous releases the change happened on the deceased strategies, it will be nice to introduce some variantions on the winning strategies too.

[ ] insert the reproduction of the winning strategies:
<25% of average points = 0 deceased
in the average = 1 survives
up to >25% = 2 survive and reproduce itself
up to >50% = 3
above = 4

[ ] the score should be modified:
example. 1 versus 2
Points 1 * Popolation 2 // Points 2 * Popolation 1
the number of games increases, but the types are the same.

[ ] Problem 1: the casual strategy will play the same for each fight against the same strategy
[ ] Problem 2: a strategy could "learn" from a first fight and act in a different manner on the games against the same strategy in the same turnament!

[ ] DElete the estinguish strategies
IF Popolation = 0 do not fight

[ ] Changes
1 every 10 (or 100)
IF Popolation <0 nothing happen
else Popolation/10 (round number) = Number mutants

Function mutante ()
mutant yes x
TOT = TOT + 1
extract DNA [x]
change it and assign the new value to DNA [TOT+1]
[ ] Problem: how to dimension the array TOT

[  ] increase the number of genes for each strategy (similar to reality)
[  ] on DNA error, random play or immediat death (as playing all 0)
[  ] decrease the number of rounds to 10

Useful references

DNA to Amino Acid Conversion: http://www.geneseo.edu/~eshamb/php/dna.php

Genetic code explanation: http://en.wikipedia.org/wiki/Genetic_code

Back to the main index of this article