Feature new DE algorithms #341
Open
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Hi, Julian!
This is the PR that includes several DE-based algorithms including fixes for compatibility purposes. Moreover, it contains the new survival operators (also included in a separate PR). You might ignore the other if this one is accepted.
I modified the previous DE file renaming it to de_ep.py and the corresponding algorithm to EPDE to preserve the parameter control strategy previously implemented. The DE basic algorithm includes now repair strategies in the native DE reproduction that can be passed as either str or callable in the de_repair argument. New parent selection schemes are also included. Dithering can be used within any range of F user-specified if F is passed as an iterable. Furthermore, jitter can be used with user-specified values for gamma.
In this PR, DEX and DES are (re)modified to be compatible with new Crossover and Selection standards from pymoo 0.6.0 but preserve the fix in degenerate DE reproduction from the older versions. Also, a class DEM is created to the DE "mutation" (or "first crossover") although it inherits from pymoo's Crossover. Both DEX and DEM are included in real crossover tests and have been verified to be compatible with GA.
The other algorithms include:
NSDE: That combined DE reproduction operators to NSGA-II survival (mu + lambda).
GDE3: Similar to NSDE but with one-to-one comparison before mu + lambda.
NSDE-R: Non-dominated Sorting Differential Evolution based on Reference directions (Reddy & Dulikravich, 2019). It combines DE with NSGA-III.
Other GDE3 variants are included as additional classes using the new crowding metrics in RankAndCrowding. They are: GDE3MNN, GDE32NN, GDE3PCD.
Feel free to reach out to me if you'd like to discuss the changes in detail.
My kind regards,
Bruno Scalia