Write a comment

Summary of the Research and Main Contributions

This preprint presents a comprehensive, data-driven framework for assessing commercial supersonic route feasibility by integrating engineering performance metrics, regulatory constraints, and economic indicators. Using a dataset of 435 global long haul routes characterized by great circle distance, over water fraction, time savings, and market demand proxies, the author formulates feasibility prediction as a supervised machine learning problem. The study trains both an interpretable Decision Tree model and a more powerful XGBoost classifier, complemented by SHAP explainability analysis.

Key findings include:

• Over water routing fraction is the dominant determinant of route feasibility, reflecting current restrictions on overland supersonic flight.

• Time savings and route distance are secondary but influential drivers.

• Market demand plays a meaningful role primarily in borderline cases.

• XGBoost achieves very strong predictive performance (ROC AUC of about 0.997) and produces a continuous feasibility score suitable for route ranking.

• The author also deploys the framework as an interactive predictive engine, demonstrating practical applicability.

This work advances the field by providing one of the most scalable and integrated frameworks for assessing route level supersonic feasibility. It helps bridge a gap in the literature between aircraft level design studies, scenario based demand modeling, and operational routing analyses.

Major Issues

1. Feasibility Labeling Criteria Lack Transparency and Validation The paper references domain informed criteria, but the labeling methodology is not described in enough detail for replication or thorough evaluation. Without explicit thresholds or sources, the labels may encode assumptions that drive the model's strong performance.

2. Regulatory Constraint Modeling Is Oversimplified Over water fraction is used as the primary regulatory proxy, but it does not distinguish between countries with different supersonic flight policies. It also does not account for potential low boom corridors, polar routing allowances, or evolving FAA and ICAO rulemaking. This limits the accuracy of feasibility predictions for certain regions.

3. Performance Assumptions Lack Sufficient Justification The paper states that representative next generation supersonic cruise assumptions were used, but it does not specify fuel burn models, Mach numbers, climb and descent profiles, or mass assumptions. These missing details make it hard to generalize findings across different aircraft concepts.

4. Economic Demand Approximation May Oversimplify Reality Using population and GDP per capita as a proxy for premium travel demand does not capture business travel patterns, airline network structure, income distribution, or competition from subsonic premium products. This may distort demand characterization for some routes.

5. Class Imbalance Raises Concerns About Model Evaluation Only 14 percent of routes are considered feasible. Although the ROC AUC score is excellent, performance on the minority class should be emphasized more (for example, precision recall curves and confusion matrices). The small number of feasible routes may inflate performance.

6. Lack of External Validation The model is not validated against historical Concorde routes, prior industry feasibility studies, or expert assessment. This limits confidence in its real world applicability.

Minor Issues

1. Clarity and Flow The introduction is lengthy and would benefit from more focused structuring. Consolidating parts of the literature review could improve flow.

2. Reproducibility Details The manuscript would benefit from explicit reporting of hyperparameters, software versions, and handling of missing data, even though the repository is provided.

3. Terminology Consistency The term feasibility score is used consistently, but the interpretation (probability vs normalized index) could be stated more clearly.

4. Data Limitations Described Briefly Although limitations are acknowledged, the discussion could quantify the reliability of population and GDP data, especially for emerging markets.

5. Formatting Issues There are small formatting inconsistencies involving spacing, table alignment, and figure placement that could be refined.

Competing interests

The author declares that they have no competing interests.

Use of Artificial Intelligence (AI)

The author declares that they did not use generative AI to come up with new ideas for their review.