RumorFlow (RF)

RumorFlow is a population-based metaheuristic inspired by rumor propagation in social networks. It models candidate solutions as "rumors" that spread and evolve on a (possibly dynamic) communication graph. The project includes a full C++ implementation, laboratory scripts, the research paper and presentation slides used in the original evaluation.

Why you might care:

• Designed for high-dimensional continuous optimization.
• Novel asymmetric propagation operator that favors high-quality solutions while keeping diversity via mutation and reinitialization strategies.
• Multi-population variant (GroupsRumorFlow) to enhance exploration and avoid premature convergence.

---

Quick links

• Paper: latex/RumorFlow.pdf or RumorFlow.pdf in repo root
• Slides: DiapositivasRumorflow.pdf
• Example entry point: main.cpp

Features

• Socially-inspired propagation and mutation operators
• Population represented as a graph (static or dynamic topology)
• GroupsRumorFlow: multi-population extension
• Benchmarking setup against CEC functions (in code/)

Repository layout

Top-level files and folders you'll likely use:

CMakeLists.txt        # CMake project file
main.cpp              # example runner / entry point
code/                 # CEC benchmark wrappers and tests
common/               # utilities, random, helpers
inc/                  # headers for RumorFlow classes
src/                  # implementation files
latex/                # paper sources
DiapositivasRumorflow.pdf
RumorFlow.pdf
input_data/           # sample datasets used in experiments
results_RumorFlow/    # output from runs and experiments

Quickstart (build & run)

Requirements: a C++17-capable compiler (g++/clang++), CMake >= 3.10, make or ninja.

Build with CMake (recommended):

git clone https://github.com/GabrielFranciscoSM/RumorFlow.git
cd RumorFlow
mkdir -p build && cd build
cmake ..
cmake --build . -- -j

Run the produced executable (from build/):

./main          # may accept command-line options; see code/main.cpp

If the project generated a build/ from a previous run, you can also use the included build/ artifacts during development.

How to use

• The main.cpp file demonstrates how to configure and run the algorithm. Inspect inc/ and src/ for configurable parameters: population size, number of iterations, problem dimension and benchmark selection.
• Input data for experiments is under input_data/ and random seeds are in datos/seeds.txt.

Reproducing experiments

1. Prepare the benchmark configuration (CEC functions are in code/).
2. Build the project as shown above.
3. Run ./main and direct outputs into results_RumorFlow/.

For plotting and post-processing, scripts plot_divergence.py and grid_search.py are available in the repo root.

Research summary

See RumorFlow.pdf for full details. Short summary:

• RF shows strong exploitation capabilities which can lead to fast convergence.
• Pure RF can lose diversity; RF-R (reinitialization) and GroupsRumorFlow (multi-population) were developed to mitigate this.
• The final GRF variant demonstrates a good trade-off between exploration and exploitation and achieves competitive results on CEC benchmarks.

Citation

If you use RumorFlow, please cite:

@misc{SanchezMunoz2025,
  author = {Gabriel Francisco Sánchez Muñoz},
  title = {RumorFlow (RF): Una Metaheurística de Inteligencia Colectiva Inspirada en la Dinámica Social},
  year = {2025},
  publisher = {GitHub},
  howpublished = {\url{https://github.com/GabrielFranciscoSM/RumorFlow}}
}

License

This project is licensed under the GNU General Public License v3.0 — see LICENSE.

Contact / Contributing

If you find issues, have suggestions or want to contribute, please open an issue or a pull request on the GitHub repository. For direct contact, check the GitHub profile of the repository owner.