Test Functions for Swarm Algorithms

6 NOT 2025

I’m setting myself the goal of working towarded a release where I an integrate a TDA submodule for dimensionality reduction.

28 OCT 2025

Today I’m adding respect to swarmopt. More specifically a respect boundary around target points. Chat suggested this would be good with satelite positioning. Don’t want to smush into something when we can give some respectful distance.

I also added this idea of sending in a few delegates within the resepct boundary, using the example of sending in repair around a discovered satelite as possible real world application.

I still want to develop my test cases out. Curious what has been developing in the literature.

26 OCT 2025

I’m interested in refactoring this project to something more elegant and seeing where PSOs can tag team with topological data analysis to accompany us in medical research. I asked ChatGPT where I could research something novel and it suggested an oncology example where I could compute barcodes of volumes and their microstructures, choose an agent-based model, and use an MPSO (which is now operational in swarmopt) to minimize the distance between simulated and observed topological summaries. I can see that. Particles dancing around the simplicial complexes and telling the story of what they see. We then extend to intervention scenarios. I’m sure people are doing that in research. I’ll read up to see what people are finding.

25 OCT 2025

Today I collaborated with Cursor to get my project swarmopt up and running as I have some fresh enthusiasm to build out some new test functions for swarm algorithms. I never think I’ll be offering something novel or game-changing in this space, but I believe in science we should get proximate to different conversations so when breakthroughs happen we can enjoy the discoveries. Insert your sports team feels, and compare that to my abiding love of the Denver Nuggets or my brother’s dedication beyond reason to the Cubs.

I am going to review my code with colleagues and speak to where I see myself in my code and where I see Cursor and how I feel about that. I’ll write about that here. I feel the scans of my code and suggesions were pretty straighforward. I thought the PR description was helpful. I want to use it as inpiration to put more care into my documentation of this project. I now want to see if I can get the utility of the project to be a bit more clear with some visualizations and test suites.

I felt like Cursor really got my project back on its feet today, and honestly I would have kicked the visualizations down the road for a long time. In the span of a handful of hours I went from a project that I’d poked at for years, mostly by gathering PSO papers and thinking of real-world problems to solve, to a project where 80% of what you see was generated by an AI assistant. It was fun to iterate and have Chat open as well where I’ve been in more dialogue about research. I want to refactor what we made today and do some writing about why I keep returning this project, as I keep saying it’s because of the metaphoric possibilities for tracking information moving in society but there’s something more. I think it’s like the question of “if you could bring one book to a desert island”, only it’s “if you could bring one handy dandy algorithm to a desert planet”.