Replace optimlib particle swarm algorithm by equivalent ROOT minimizer.

[FonsRademakers]

Is your feature request related to a problem? Please describe.
Try to minimize the number of dependencies as optimlib also pulls in lapack and cblas.

Describe the solution you'd like
Try to replace the optimlib Particle Swarm Optimization with an appropriate ROOT minimizer.

[vasvav]

A couple of general questions and suggestions (no need to answer here):

• The reported issue, is it so very important to resolve? Is there a general consensus to put development effort in this new task? Has this topic being discussed in a technical meeting?
• This issue sounds as a wish-to-have item, not an issue to report fixing, correct?
• In general, it is good practice that proposals for prospective features / API developments should be discussed during BioDynaMo technical telcos, not through GitHub please.

[FonsRademakers]

Hi Vas,

I had this for a while on my wish list. The issue is that if we can reduce complexity we should try to do that. Also there were a number of installation failures related to optimlib versions. The more you led this kind of issues pass without addressing them the messier the code becomes.

I created this issue as a reminder that we might spend sometime some time on seeing if this can be fixed using a library we already use. If the conclusion is that is not possible then we can close this issue. Maybe Ahmad can look into it as he introduced this code. It is marked as a feature request and not as a bug.

I think people should be able to propose API changes and new features via Github issues, not all people with good ideas participate in technical telcos.

[TobiasDuswald]

In general, it is good practice that proposals for prospective features / API developments should be discussed during BioDynaMo technical telcos, not through GitHub please

I think people should be able to propose API changes and new features via Github issues, not all people with good ideas participate in technical telcos.

Briefly wanted to comment on the above. I think @FonsRademakers has a very good point. I'd generally advocate for making the development as public and transparent as possible so that people can engage in the development, get inspiration, and contribute ideas and visions to the project. GitHub issues are a perfect tool for that. If an issue is created for a feature, people can start picking it up and supporting it if it's important to them. It also shows a bit of a heartbeat; the project is very much alive. :)

Thought that's worth commenting on. But I don't wanna steer the conversation in the wrong direction. Let's get back to the optimlib topic for which this feature request was created ;)

[romanbauer111]

Thanks for all your comments, this is an important discussion. I agree we should have a transparent way to propose changes and feature requests. But we also need to find a way to agree on these changes and their priorities, and to assign people. Also we should have differential channels for internals and externals. Otherwise it becomes chaotic. That is one of the main reasons for having a Discord channel, so that we can streamline discussions and eventually agree internally. Let us discuss this in more detail at the next BioDynaMo meeting.