Radio Resource Allocation for Beam Hopping Scheduling in LEO Satellite Communications: A Spatio-Temporal Perspective

⚠️ Important Update (2026-03-11): All code errors have been fixed!

If you cloned this repository before, please update immediately:

git pull origin main

See Changelog for details

---

This repository contains the MATLAB implementation of the Radio Resource Allocation for Beam Hopping Scheduling in LEO Satellite Communications: A Spatio-Temporal Perspective.

📋 Overview

This project addresses the beam-hopping (BH) scheduling problem in LEO satellite communications, where a satellite dynamically illuminates multiple cells within its coverage area to maximize service satisfaction while managing interference constraints.

Key Features

• Tabu Search with Simulated Annealing (SA): Hybrid metaheuristic combining tabu search's memory mechanism with SA's probabilistic acceptance
• Adaptive Tabu Tenure: Dynamic tabu list size based on problem scale ($L_{tabu} = \lfloor\sqrt{N_b} \cdot \sqrt{N_m}\rfloor$)
• Interference-Aware Scheduling: Spatial separation constraints to minimize co-channel interference
• User-Centric KPIs: Comprehensive performance metrics including throughput percentiles, fairness, and service satisfaction

🚀 Quick Start

Prerequisites

• MATLAB R2024a or later
• Required toolboxes:
  • Communications Toolbox
  • Optimization Toolbox (optional, for baseline comparison)
• Satellite Orbit Data: This project requires satellite orbit data files (5400.mat or 1800.mat) containing position information. Run generate_test_satellite_data() to generate sample data.

Installation

# Clone the repository
git clone https://github.com/yuanhaobupt/leo-bh-scheduling.git

# Navigate to the project directory
cd leo-bh-scheduling

# Open MATLAB and add to path
addpath(genpath('.'));

Running a Simple Example

% Load default configuration
setConfig;

% Run a quick test (single scheduling period)
controller = simSatSysClass.simController(Config, 1, 1, 0);
DataObj = controller.run();

% Calculate performance metrics
KPIs = calcuUserKPIs(DataObj);

📁 Repository Structure

leo-bh-scheduling/
│
├── +methods/                    # Algorithm implementations
│   ├── BHST_MY.m               # Original Tabu Search
│   ├── BHST_MY_SA.m            # Tabu Search with SA
│   ├── BHST_greedy.m           # Greedy baseline
│   ├── BHST_DQN.m              # DQN baseline
│   └── ...
│
├── +simSatSysClass/            # Simulation framework
│   ├── @simController/         # Main controller
│   ├── @simInterface/          # Data interface
│   ├── @dataObj/               # Result storage
│   └── +tools/                 # Utility functions
│
├── +antenna/                   # Antenna pattern functions
├── +tools/                     # Coordinate and geometry utilities
│
├── experiments/                # Experiment scripts
│   ├── run_TabuSearch.m        # Tabu Search baseline
│   ├── run_DQN.m               # DQN baseline
│   ├── run_ablation_SA.m       # SA ablation experiment
│   ├── run_ablation_Ltabu.m    # Tabu tenure ablation
│   ├── run_traffic_skew_experiment.m  # Traffic skew experiment
│   ├── run_all_experiments.m   # Comprehensive runner
│   └── run_visualization.m     # Batch visualization
│
├── utils/                      # Helper utilities
│   ├── calcuUserKPIs.m         # User-centric KPI calculation
│   ├── generateTraffic.m       # Traffic demand generation
│   ├── generate_test_satellite_data.m  # Synthetic orbit data
│   ├── analyze_existing_results.m      # Result analysis
│   └── perform_statistics.m    # Statistical analysis
│
├── visualize/                  # Visualization tools
│   ├── plot_SINR_CDF.m
│   ├── plot_Throughput_Comparison.m
│   ├── plot_ablation_results.m
│   └── ...
│
├── results/                    # Experiment results (git-ignored)
│
├── setConfig.m                 # Configuration script
├── quick_start.m               # One-click verification
│
└── README.md                   # This file

🔬 Experiments

1. SA Mechanism Ablation

Compare the proposed method (Tabu + SA) against Tabu-only baseline:

run('experiments/run_ablation_SA.m')

Expected Results:

Variant	Throughput (Mbps)	Satisfaction (%)
Tabu + SA	202.6	85.0
Tabu-only	189.3	82.5

2. Tabu Tenure Ablation

Compare adaptive vs. fixed tabu tenure:

run('experiments/run_ablation_Ltabu.m')

Expected Results:

Configuration	Throughput (Mbps)	Satisfaction (%)
Fixed L=10	178.5	79.8
Fixed L=20	195.2	82.5
Fixed L=30	188.3	81.2
Adaptive	202.6	85.0

3. Traffic Skew Experiment

Test performance under different traffic distributions:

run('experiments/run_traffic_skew_experiment.m')

Traffic Modes:

• Uniform: All users have identical demands
• Light Skew: 2× demand variance
• Heavy Skew: 5× demand variance
• Pareto (80/20): 20% users account for 80% demand

⚙️ Configuration

Key parameters can be modified in setConfig.m:

% Satellite parameters
heightOfSat = 508e3;           % Orbital altitude (m)
numOfServbeam = 10;            % Number of beams

% Scheduling parameters
BhDispCycle = 40e-3;           % Scheduling period (s)
SubCarrierSpace = 30e3;        % Subcarrier spacing (Hz)

% Traffic parameters
meanUsrsNum = 800;             % Average number of users
traffic_mode = 'uniform';      % Traffic distribution mode

% Algorithm parameters
enable_SA = true;              % Enable SA mechanism
L_tabu_mode = 'adaptive';      % Tabu tenure mode

📊 Performance Metrics

The framework calculates the following user-centric KPIs:

Metric	Description
Throughput	Average, p50, p90, p95 percentiles
SINR	Signal-to-Interference-plus-Noise Ratio
Outage Rate	Users with SINR < 0 dB
Service Satisfaction	Transported/Requested traffic ratio
Fairness Index	Jain's Fairness Index
Delay	Average and p95 latency

Example usage:

KPIs = calcuUserKPIs(DataObj);
fprintf('Average Throughput: %.2f Mbps\n', KPIs.avg_throughput/1e6);
fprintf('p90 Throughput: %.2f Mbps\n', KPIs.p90_throughput/1e6);
fprintf('Fairness Index: %.4f\n', KPIs.fairness_index);

📈 Complexity Analysis

Time Complexity (per iteration): $$T_{iter} = O(|N(x)| \cdot N_b \cdot K)$$

Where:

• $|N(x)|$: Neighborhood size (default: 10)
• $N_b$: Number of beams (default: 10)
• $K$: Average users per beam (default: 80)

Space Complexity: $$S_{total} = O(L_{tabu} \cdot N_b + N_m^2)$$

Typical Performance:

• Computation time: ~5 seconds (50 iterations)
• Memory usage: <100 MB
• Suitable for real-time satellite scheduling

📝 Citation

Waiting...

📄 License

This project is licensed under the MIT License - see the LICENSE file for details.

🤝 Contributing

We welcome contributions! Please see CONTRIBUTING.md for guidelines.

📧 Contact

For questions or issues:

• Hao Yuan - Email

---

Note: This code is provided for academic research purposes. For commercial use, please contact the authors.