m-decoster / rsgenetic Goto Github PK
View Code? Open in Web Editor NEWGenetic Algorithm library in Rust
License: Apache License 2.0
Genetic Algorithm library in Rust
License: Apache License 2.0
Work has started on this in the less-copy
branch.
Hi!
Thanks for the nice library!
I'd like to get some clarification about Fitness
abstraction. What it exists for, and why it can not be just f64
? (So Phenotype's fitness
function returns only f64
). The smaller(greater) returned fitness, the better are parameters.
I guess there are some ideas behind Fitness
, that I could not get.
Another question: Does the algorithm finds the smallest fitness or the biggest one?
Thanks in advance!
P.S. Please consider adding this info to README, maybe I'm not the one who's confused :)
All selection algorithms should return count
parents. This requires a rewrite of some selection algorithms.
An explicit get
function should be added to avoid users bypassing errors. The run
and step
functions should not return the intermediate results anymore.
This should be an option.
Reference. The O(1) algorithm that is mentioned on that page should be implemented, because it's fast and easy.
I profiled a benchmark just now and noticed that 99% of CPU time was spent sorting in the maximize selector. The problem is so bad that the execution time difference between a sequential and parallel simulator is very minimal.
Here are execution times (in nanoseconds) by selector types for a sequential Simulator for the benchmark in separate_points
:
Parameter | Maximize | Stochastic | Tournament |
---|---|---|---|
Execution time (ns) | 26.4 * 10^9 | 24.2 * 10^4 | 10.1 * 10^7 |
Parameters | 10 | 10 | 10, 5 |
This large difference may be an issue with regards to execution time.
There should be at least a coarse-grain parallel implementation for the Simulator
.
This would also require a refactoring of the sim
module.
The current SimulatorBuilder
is implemented in a consuming builder pattern. According to https://aturon.github.io/ownership/builders.html the preferred is the non-consuming one.
IMO the non-consuming pattern may be better here:
set_*
functions implies it is mutable and does not return new instance.set_*
however assigning to a new builder requires more word typings.This example could be the same as from the documentation.
I suggest implementing the Builder pattern for Simulation, so that we can update the minor versions without breaking backwards compatibility. Having to update the constructors would break this.
I suggest adding a function to the Simulation
trait that returns a builder specific to the implementation.
This can be useful to compare the impact of certain Simulation
implementations or different SelectionType
on performance.
I suggest letting the fn run(&mut self)
function, which currently returns ()
, return i64
instead, containing the time passed in this function. The new definition would be fn run(&mut self) -> i64;
.
An alternative is adding a seperate time
function, but this would require implementors of this trait to store more fields, introducing more state. I'd like to keep state to a minimum.
I am creating this issue to ask potential users' input on a question.
For version 1.0.0, I am considering making Fitness
a trait, and making Phenotype
generic over F: Fitness
. This would allow users of the library to define their own fitness types, and not have to make the mapping to f64
, which might be troublesome in some cases.
This feature would define a trait pub trait Fitness : Ord + Eq
to be implemented and passed to Phenotype
, Simulation
and Selector
.
With this feature, the library would become almost completely generic:
However, this feature would introduce some extra development time up front, because the Fitness
trait would have to be implemented.
Finally, this change implies the removal of FitnessType
, because this can be implemented by the user.
Hi! This might be a bit of a weird one, but it would be great to have WASM support for this crate ๐
... which I have gotten to work, but now I am not sure how to contribute back to you guys :) there are two major changes that I had to implement:
Now I'd be happy to integrate this somehow, but I am not sure how to do this without rewriting half the seq module or removing things like the timing and replacing thread_rng (where I don't know the difference in quality to XorShiftRng). Alternatively it could live as its own compilation unit for the wasm target, but that would duplicate code which is also not great ๐
What are your ideas? I am happy to help ๐
This enables reuse of selectors, which will be useful for #1. We can also move the selector implementation to a separate module, reducing the bloat in sim::seq
.
A step
function for Simulation
s which only makes a single step could be useful for Simulation
objects, to make a visualization of the evolution of organisms.
We need to add some checks to avoid this. These checks need to be added in the selection_*
functions in src/sim/seq.rs
.
Hi there.
It would be nice to have pluggable crossover and mutation operators and implementations of well known operators.
This issue was automatically generated. Feel free to close without ceremony if
you do not agree with re-licensing or if it is not possible for other reasons.
Respond to @cmr with any questions or concerns, or pop over to
#rust-offtopic
on IRC to discuss.
You're receiving this because someone (perhaps the project maintainer)
published a crates.io package with the license as "MIT" xor "Apache-2.0" and
the repository field pointing here.
TL;DR the Rust ecosystem is largely Apache-2.0. Being available under that
license is good for interoperation. The MIT license as an add-on can be nice
for GPLv2 projects to use your code.
The MIT license requires reproducing countless copies of the same copyright
header with different names in the copyright field, for every MIT library in
use. The Apache license does not have this drawback. However, this is not the
primary motivation for me creating these issues. The Apache license also has
protections from patent trolls and an explicit contribution licensing clause.
However, the Apache license is incompatible with GPLv2. This is why Rust is
dual-licensed as MIT/Apache (the "primary" license being Apache, MIT only for
GPLv2 compat), and doing so would be wise for this project. This also makes
this crate suitable for inclusion and unrestricted sharing in the Rust
standard distribution and other projects using dual MIT/Apache, such as my
personal ulterior motive, the Robigalia project.
Some ask, "Does this really apply to binary redistributions? Does MIT really
require reproducing the whole thing?" I'm not a lawyer, and I can't give legal
advice, but some Google Android apps include open source attributions using
this interpretation. Others also agree with
it.
But, again, the copyright notice redistribution is not the primary motivation
for the dual-licensing. It's stronger protections to licensees and better
interoperation with the wider Rust ecosystem.
To do this, get explicit approval from each contributor of copyrightable work
(as not all contributions qualify for copyright, due to not being a "creative
work", e.g. a typo fix) and then add the following to your README:
## License
Licensed under either of
* Apache License, Version 2.0 ([LICENSE-APACHE](LICENSE-APACHE) or http://www.apache.org/licenses/LICENSE-2.0)
* MIT license ([LICENSE-MIT](LICENSE-MIT) or http://opensource.org/licenses/MIT)
at your option.
### Contribution
Unless you explicitly state otherwise, any contribution intentionally submitted
for inclusion in the work by you, as defined in the Apache-2.0 license, shall be dual licensed as above, without any
additional terms or conditions.
and in your license headers, if you have them, use the following boilerplate
(based on that used in Rust):
// Copyright 2016 RsGenetic developers
//
// Licensed under the Apache License, Version 2.0 <LICENSE-APACHE or
// http://www.apache.org/licenses/LICENSE-2.0> or the MIT license
// <LICENSE-MIT or http://opensource.org/licenses/MIT>, at your
// option. This file may not be copied, modified, or distributed
// except according to those terms.
It's commonly asked whether license headers are required. I'm not comfortable
making an official recommendation either way, but the Apache license
recommends it in their appendix on how to use the license.
Be sure to add the relevant LICENSE-{MIT,APACHE}
files. You can copy these
from the Rust repo for a plain-text
version.
And don't forget to update the license
metadata in your Cargo.toml
to:
license = "MIT/Apache-2.0"
I'll be going through projects which agree to be relicensed and have approval
by the necessary contributors and doing this changes, so feel free to leave
the heavy lifting to me!
To agree to relicensing, comment with :
I license past and future contributions under the dual MIT/Apache-2.0 license, allowing licensees to chose either at their option.
Or, if you're a contributor, you can check the box in this repo next to your
name. My scripts will pick this exact phrase up and check your checkbox, but
I'll come through and manually review this issue later as well.
https://github.com/m-decoster/RsGenetic/blob/master/src/sim/select/tournament.rs#L63-L71
For tournament selection, line 70 requires that at least there are 2 elements; when 1 participant is specified in user code, it will cause index out of bound panic.
Currently, the crossover
method can only return one offspring.
But often it is desirable to have high fitness parents producing more offspring.
And it's not just enough to call the crossover
function multiple times, because often the offspring should be generated exhaustively, regarding a certain set of "orthogonal" crossover combinations that makes sense, without repeating the same way of crossover.
Simply calling crossover
multiple times would potentially produce the same offspring multiple times, if the number of crossover-ways that make sense is low.
The same can be said for the mutate
method. Often it's desirable to generate multiple "orthogonal" mutations together.
A declarative, efficient, and flexible JavaScript library for building user interfaces.
๐ Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. ๐๐๐
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google โค๏ธ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.