Add NSGA2

[say4n]

This PR implements the Non-dominated Sorting Genetic Algorithm - II by Kalyanmoy Deb.
It is a Work in Progress. :)

TODO:

• Implement NSGA-II
• Add tests
• Add documentation

[favre49]

I've only made a quick read of it, I'm currently on vacation so I didn't get much time to read it deeply. In the meantime, could you document the code too? That could aid the reviewers in understanding what your code is doing and make this process quicker :)

[say4n]

In the meantime, could you document the code too?

Yes, will do it ASAP! :)

[rcurtin]

I installed valgrind on bigglesworth. 👍 Thanks so much @say4n, great work. I'll approve once I see it pass all the builds.

[say4n]

@rcurtin there are some issues with the memcheck CI job. Marcus said he'll look into it (on IRC).

The remainder of the CI jobs are green.

[zoq]

Valgrind over all tests:

real    553m53.868s
user    553m21.108s
sys     0m23.900s

<?xml version="1.0"?>

<valgrindoutput>

<protocolversion>4</protocolversion>
<protocoltool>memcheck</protocoltool>

<preamble>
  <line>Memcheck, a memory error detector</line>
  <line>Copyright (C) 2002-2017, and GNU GPL'd, by Julian Seward et al.</line>
  <line>Using Valgrind-3.13.0 and LibVEX; rerun with -h for copyright info</line>
  <line>Command: build/ensmallen_tests -s</line>
</preamble>

<pid>1819</pid>
<ppid>1815</ppid>
<tool>memcheck</tool>

<args>
  <vargv>
    <exe>/usr/bin/valgrind.bin</exe>
    <arg>--tool=memcheck</arg>
    <arg>--leak-check=full</arg>
    <arg>--track-origins=yes</arg>
    <arg>--num-callers=40</arg>
    <arg>--error-exitcode=1</arg>
    <arg>--xml=yes</arg>
    <arg>--xml-file=reports/tests/ensmallen.xml</arg>
    <arg>--log-file=reports/tests/ensmallen.memcheck</arg>
    <arg>--verbose</arg>
  </vargv>
  <argv>
    <exe>build/ensmallen_tests</exe>
    <arg>-s</arg>
  </argv>
</args>

<status>
  <state>RUNNING</state>
  <time>00:00:00:00.141 </time>
</status>


<status>
  <state>FINISHED</state>
  <time>00:09:13:53.384 </time>
</status>

<errorcounts>
</errorcounts>

<suppcounts>
</suppcounts>

</valgrindoutput>

Looking at the runtime, I'm going to modify the mem-check job to run only related tests.

[say4n]

Whoa ~10 hours!

[rcurtin]

Hey @favre49, do you want to take another pass at a review on this? I think it's basically ready. 👍

[favre49]

Glad to see this completed, no comments from my side. Great work! :)

[favre49]

I think once we merge this we can close #120, since this basically invalidates it. If I get the time again I can reopen it to conform to this interface and such.

[rcurtin]

@favre49 agreed, that sounds good. If you like, maybe it is easy to adapt some small parts of that PR.

I guess we should wait to merge this one until the memory check job is resolved?

[zoq]

I can put some time to resolve the memory job runtime issue, unless @say4n already is close to push a patch?

[say4n]

@zoq I haven't yet started working on it, I was planning to do it in the weekend ahead. 😛

[zoq]

Okay, I went ahead and put something together: mlpack/jenkins-conf#23

[rcurtin]

Awesome, mlpack/jenkins-conf#23 is merged now. I think maybe some Jenkins configuration changes might still be needed before re-running though?

[zoq]

Okay, I think I have everything set up correctly now.

[zoq]

@mlpack-jenkins test this please

[say4n]

Whoa, the memory checks ran super fast!

[rcurtin]

Looks like maybe it didn't run any tests?

<testsuite name="ensmallen_tests" errors="0" failures="0" tests="0" hostname="tbd" time="0.141205" timestamp="2020-08-09T21:24:36Z">

I might have read the output wrong though.

[zoq]

@mlpack-jenkins test this please

[zoq]

Looks like maybe it didn't run any tests?

<testsuite name="ensmallen_tests" errors="0" failures="0" tests="0" hostname="tbd" time="0.141205" timestamp="2020-08-09T21:24:36Z">

I might have read the output wrong though.

No, let's see if I got it right this time.

[say4n]

So much green. :)

[rcurtin]

Awesome! It looks good to me and the memory check scripts seem to work fine. I'll go ahead and merge it.

@say4n do you want to help with the release? If you're willing, try this:

• check out the ensmallen master branch
• from the root of the repository, run scripts/ensmallen-release.sh say4n 2 14 0 "<name>" where <name> is whatever you think the next release name should be :)

I'm not sure it will work completely, but if it does, then it will automatically open a pull request using the hub tool for the next release. Anyway, I can also do it, just let me know.

[say4n]

Sure, I can create the PR for a new release.

Do we keep it Automatically Automated Automation? If not, does the name need to pertain to a certain theme?

[rcurtin]

Nah, for a new minor release, feel free to pick a new name. :)

[say4n]

Sure, on it. :P

[rcurtin]

If you have any problems just let me know---the script is still a work in progress. 😄

[say4n]

Ah, so I've opened #214 but the changelog in the description for the PR is was wrong (I've corrected it now).