Skip to content

About

Multi-objective evolutionary routing optimization for networks: NSGA-II, SPEA2 and INSGA in C++, with a Python CLI for experiments.

Resources

Stars

0 stars

Watchers

0 watching

Forks

Repository files navigation

multi-objective-evolutionary-network-routing

Multi-objective evolutionary algorithms for routing in telecommunication networks. Given a network topology (e.g. from SNDlib) and a set of traffic intents (demands between nodes), the project searches for sets of paths that trade off throughput, packet loss, delay and jitter, and returns a Pareto front of optimal solutions.

  • Algorithms (C++): SPEA2, INSGA, and NSGA-II.
  • CLI (Python): parses SNDlib networks, generates full-mesh topologies and intents, approximates true Pareto fronts, compares the algorithms, and tests metric convergence and computational complexity, with plots.

Table of contents

  1. Project Requirements
  2. Branch Structure
  3. Source Code Navigation

How to run CLI client?

Setup:

Create python environment with required libraries:

python -m venv .venv/
source .venv/bin/activate
pip install -r ./src/requirements.txt

Build required C++ programs (make sure you are in project folder)

cmake -S . -B build && cmake --build build

Usage:

Open project folder and run run.py as module:

cd <project folder>
python -m src.run [subcommand] [subcommand options and flags]

For help and additional information run:

python -m src.run -h
# or for subcommand
python -m src.run [subcommand] -h

Example Use Cases

Graph Generation

  1. Parsing graph from sndlib
# python -m src.run parse-sndlib network_name
python -m src.run parse-sndlib polska
  1. Generating fullmesh graph
# python -m src.run gen-full-mesh node_count
python -m src.run gen-full-mesh 20

Intents Generation

# python -m src.run gen-intents graph
python -m src.run gen-intents polska.csv

True Pareto Front Approximation Generation

# python -m src.run gen-true-pareto graph intents iterations runs mutation_probability
python -m src.run gen-true-pareto polska.csv polska_intents.csv 1000 20 0.4

Algorithm Comparison

# python -m src.run algo-compare graph intents true_pareto_frton iterations runs mutation_probability --plot-data(optiona flag)
python -m src.run algo-compare polska.csv polska_intents.csv polsk_tpf.csv 1000 15 0.4 --plot-data

Computational Complexity Tests

# python -m src.run c-comp [--load-data FILENAME] [--iterations ITERATIONS] [--runs RUNS] [--mutation MUTATION] [--min_nodes MIN_NODES] [--max_nodes MAX_NODES] [--step STEP]
python -m src.run c-comp --iterations 100 --runs 1 --mutation 0.1 --min_nodes 5 --max_nodes 20 --step 2

Convergence of Metrics

# python -m src.run check-conv [--plot-data] graph intents true_pareto iterations runs mutation_probability max_population step
python -m src.run check-conv polska.csv polska_intents.csv polsk_tpf.csv 1000 15 0.4 30 10 --plot-data

About

Multi-objective evolutionary routing optimization for networks: NSGA-II, SPEA2 and INSGA in C++, with a Python CLI for experiments.

Resources

Stars

0 stars

Watchers

0 watching

Forks

Releases

Contributors

Languages