ASPaeroFlow-Optimizer

ASPaeroFlow-Optimizer is a research codebase for solving the Strategic Air Traffic Flow and Capacity Management (ATFCM) problem with a combination of Answer Set Programming (ASP) and heuristic optimization. It provides algorithms to jointly optimize air traffic flow management measures (ground/en-route delays and reroutes) together with Dynamic Airspace Configuration (DAC) decisions (sector capacity adjustments) under a unified model. The repository includes a novel ASPaeroFlow heuristic that iteratively resolves sector overloads using ASP sub-solvers, as well as exact and baseline solvers for comparison.

Features

• Simultaneous Optimization: Jointly optimizes flow actions (delay, rerouting) and airspace reconfiguration (DAC)
• Instance-Space Decomposition: Scalable heuristic that solves local overloads with ASP
• ASP-Based Core Solver: Lexicographic soft objectives (safety, delay, stability, changes)
• Open Data Compatibility: Accepts input from generator based on OpenSky Network data
• Baselines Included: MILP solver, delay-only heuristic, exact ASP model for small instances
• Benchmarking Tools: Batch runner and CSV output for experiments

Repository Structure

├── 01_ASPaeroFlow/              # Main heuristic solver
├── 02_ASP/                      # Exact ASP solver (small instances)
├── 04_MIP/                      # MILP solver baseline (Gurobi)
├── 06_benchmark_start_script/   # Scripts to benchmark across solvers

Installation

Requirements:

• Python 3.9+
• Dependencies: numpy, pandas, scipy, networkx, joblib, psutil, clingo
• ASP solver: Clingo
• (Optional) Gurobi for MIP solver (gurobipy)

Install with pip:

pip install numpy pandas scipy networkx joblib psutil clingo

Ensure clingo is available via the Python API or command line.

Usage

Ensure you have an example in example_instance from ASPaeroFlow-DataGenerator.

Run ASPaeroFlow Heuristic

python 01_ASPaeroFlow/main.py \
  --data-dir example_instance
  --encoding-path 01_ASPaeroFlow/encoding.lp \
  --wandb-enabled=false \
  --verbosity 1

Output shows overload resolution progress and final objective.

Run Exact ASP Solver (Small Instances)

python 02_ASP/main.py \
  --data-dir example_instance
  --encoding-path 02_ASP/encoding.lp --verbosity 1

Prints delay, reroutes, and final schedule.

Run Baselines

• MILP (flow-only): 04_MIP/main.py (requires Gurobi)

Batch Benchmarking

Provided in 06_benchmark_start_script/start_benchmarks.py Generates performance CSVs per solver and instance.

Citation

License

MIT license with attribution (for details see license.md).

Contributions

We welcome academic contributions and extensions. Please contact the authors for collaboration or open a pull request.

---

For questions or collaborations, please reach out to the original authors. This README was created with the help of generative AI.