-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathChangeLog
36 lines (22 loc) · 2.16 KB
/
ChangeLog
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
2024-07-20 Giovanny Covarrubias Pazaran <[email protected]>
* [r1] Initial release: brining a simple genetic algorithm for complex problems.
2024-08-20 Giovanny Covarrubias Pazaran <[email protected]>
* [r2] Improve to have a single fitness function across traits. More efficient matrix operations. Additional examples for how to do regression and add contraints.
2024-09-25 Giovanny Covarrubias Pazaran <[email protected]>
* [r2] Added additional arguments for stopping criteria and to keepBest to trace the path to the best solution.
* [r2] Added mutRate and nChr arguments to control the mutation rate and recombination rate.
2024-10-15 Giovanny Covarrubias Pazaran <[email protected]>
* [r3] Now the traits with zero variance also work without errors
* [r3] We print in the console the generation, number of removed solutions/individuals based on constraints, total number of individuals left and genetic variance
* [r3] Now we allow to set different number of chromosomes which allows for a greater recombination rate by avoiding the LD and interference of having too many QTLs in the same chromosome
2025-01-18 Giovanny Covarrubias Pazaran <[email protected]>
* [r4] We set the OMP_THREAD_LIMITS to 2 with the .onLoad function to avoid CRAN checks to reject for using more than 2 cores during the vignette building step
2025-04-18 Giovanny Covarrubias Pazaran <[email protected]>
* [r5] Now the returning object from the evolafit function is an extended class of the 'Pop-class' object from AlphaSimR, an S4 object.
* [r5] We changed the name of the argument to make the use of fitness functions more straightforward.
* [r5] We added the fixQTLperInd argument to allow users to fix the number of activated QTLs for regression optimization problems
* [r5] A fitness function named regFun has been added for regression problems.
## TO DO
+ Rename score by fitness in the score table
+ Double check that mutation goes both directions, maybe an additional argument.
+ Not only allow the posibility of mutation in the genome, but also in the addtive effects (mutRateTrait)