Syllabus

1. ML-Powered Traffic Forecasting in Optical Backbone Networks
Objective: Use time-series ML models (e.g., LSTM, Autoformer) to forecast optical traffic at link or node level.
Data: Synthetic traffic traces or real datasets (e.g., GEANT, Abilene).
Outcome: Builds foundation for intelligent optical network control, important in vendor automation platforms (e.g., Ciena Blue Planet, Nokia NSP).

2. Simulation of Quantum Key Distribution over Optical Networks
Objective: Simulate a QKD-enabled optical network using trusted repeater placement and analyze key generation rate and security margins for different topologies (e.g., ring, mesh).
Add-on: Students can explore how path selection and node trust levels affect secure communication range and reliability.
Outcome: Foundational knowledge for students interested in quantum communication or network security.

3. End-to-End Design and Analysis of Multi-Wavelength Optical Systems
Objective: Design a WDM system (4/8 channels) over long-haul links (e.g., 1000 km) using cascaded EDFAs in OptiSystem or MATLAB.
Task: Simulate BER performance and eye diagrams under various input power levels, spacing, and amplifier gain settings.
Outcome: Students gain insight into system-level optical performance metrics and amplifier design.

4. Optical Power Equalization Using Adaptive Control Algorithms
Objective: Simulate a WDM system with different input power levels and apply adaptive attenuation algorithms to equalize received power across wavelengths.
Add-on: Incorporate ML regression models to estimate optimal attenuation values based on channel characteristics.
Outcome: Understanding of real-world challenges like channel balancing and power tilting in DWDM systems.

5. Attenuation and Dispersion Characterization in Optical Fibers
Objective: Simulate signal attenuation and chromatic dispersion over varying fiber lengths (e.g., 0–100 km).
Task: Analyze the impact on pulse broadening and signal integrity.
Outcome: Students develop a clear understanding of link budgeting and the physical limitations of fiber.

6. Optical Network Fault Detection using Anomaly Detection Models
Objective: Generate synthetic traffic and simulate optical signal parameters (e.g., OSNR, BER).
Train unsupervised ML models (e.g., Isolation Forest, Autoencoders) to detect faults such as link failure or amplifier misalignment.
Outcome: Introduces ML-based proactive network management and prepares students for SDN/AI integration in optical networks.

To apply machine learning techniques for performance prediction, resource allocation, and fault detection in optical networks.

To model and analyze long-haul optical links incorporating cascaded EDFAs, dispersion effects, and link impairments.

To design and simulate optical WDM systems, including channel setup, multiplexing, and end-to-end signal evaluation using software tools.

1. CO1: Design and simulate WDM and elastic optical systems, including amplifiers and dispersion models, for long-distance optical communication.
2. CO2: Apply machine learning algorithms for key optical network tasks such as traffic forecasting, QoT estimation, and fault detection.
3. CO3: Analyze and optimize resource allocation, spectrum management, and energy efficiency in WDM/EON-based optical networks using heuristic and ML approaches.
4. CO4: Simulate and evaluate quantum key distribution (QKD) in optical networks through trusted repeater placement and secure routing strategies.
5. CO5: Demonstrate the use of SDN principles for dynamic control, slicing, and provisioning in simulated optical transport networks.

Djafar Mynbaev, Lowell L. Scheiner, Fiber-optic Communications Technology, Pearson College Div (January 1, 2001)

John M. Senior, Optical Fiber Communications: Principles and Practice,, Pearson Education Third edition (1 January 2014)

Zang, Hui, Jason P. Jue, and Biswanath Mukherjee. "A review of routing and wavelength assignment approaches for wavelength-routed optical WDM networks." Optical networks magazine 1, no. 1 (2000): 47-60.