Simulated Annealing (SA) has been initially proposed in Optimization by Simulated Annealing as an optimization heuristic. Multi-objective Simulated Annealing (MOSA) extends the original, single-objective SA to approximate the Pareto front in multi-objective optimization problems.
A comprehensive discussion on MOSA and its algorithm variants can be found in Multi-objective Simulated Annealing: Principles and Algorithm Variants.
If you have any questions, corrections, comments or suggestions, just drop a message.
You can also reach me on Linkedin or follow me on X.
If you want to support this and other open source projects that I maintain, become a sponsor on Github.
The easiest way to install MOSA is using pip:
pip install mosa
You can access the API documentation for MOSA on the project's GitHub Pages site.
MOSA includes portable Agent Skills for Codex, Claude Code, and other compatible agents:
- mosa-optimize: implement and tune continuous, discrete, and mixed optimization problems.
- mosa-analyze: merge archives, filter Pareto solutions, rank with TOPSIS, and plot objectives.
Discover and install the skills from GitHub using the Skills CLI (requires Node.js/npm):
npx skills add https://github.com/rgaveiga/mosa --list
npx skills add https://github.com/rgaveiga/mosa --skill mosa-optimize mosa-analyze --agent codex claude-code --copy
Add --global to the installation command for user-wide installation.
Contributions are definitely welcome. However, it should be mentioned that this repository uses poetry as a package manager.
Source code must be formatted using black.
The code is provided "as is," with no guarantees regarding the accuracy of its results. The author assumes no responsibility for any losses arising from the use of the code.
Bugs must be reported as issues on the project's GitHub repository.