This project explores the resilience and methodological limitations of London's Underground network by evaluating its topological resilience and incorporating passenger flow data to assess network vulnerability under hypothetical scenarios. The primary focus is to understand the impact of removing key stations on the network's functionality and to simulate changes in job availability and transport costs.
Topological Network Analysis
Centrality Measures
To identify key stations within the London Underground network, three centrality measures were employed: Betweenness, Degree, and Closeness centrality.
Betweenness Centrality: This measure identifies stations that act as critical bridges within the network. Stations with high betweenness centrality scores, such as Stratford and Bank and Monument, are crucial for maintaining connectivity.
Degree Centrality: This measure highlights stations with a high number of direct connections. Stratford and Bank and Monument again rank highly, indicating their roles as major hubs for passenger transfers.
Closeness Centrality: This measure identifies stations that offer the shortest average journey time to all other stations. Green Park and Bank and Monument were among the highest-ranked, suggesting their central importance in the network.
Impact Measures
The study measured the impact of station removal using Global Efficiency and Average Shortest Path Length (ASPL). These metrics provide a comprehensive view of the network's robustness and connectivity.
Node Removal Analysis
The analysis included both sequential and non-sequential node removal strategies:
Sequential Removal: Stations were removed one by one based on their centrality scores, recalculating the centrality after each removal.
Non-Sequential Removal: Stations were randomly removed from the top 10 list, providing a more realistic simulation of unexpected disruptions.
Results indicated that Betweenness centrality was the most effective measure in determining network vulnerability, with its node removals causing significant disconnections.
Weighted Network Analysis
Incorporating passenger flow data provided a more nuanced understanding of network resilience.
Flow-Weighted Betweenness Centrality
This adjusted measure accounted for the volume of passengers passing through each station. Key stations identified included Bank and Monument and King's Cross St. Pancras, reflecting their importance in managing passenger flows.
Impact Measures Transformation
Global efficiency was recalculated to account for weighted path lengths, providing insights into how well the network facilitates passenger transfer under varying conditions.
Node Removal in a Weighted Network
Sequential removal of top nodes based on flow-weighted Betweenness centrality revealed that stations like Canada Water had a significant impact on network efficiency when removed.
Spatial Interaction Models
The project also employed spatial interaction models to simulate the effects of changes in job availability and transport costs on passenger flows.
Model Selection and Calibration
An Origin Constrained Model with exponential decay and Poisson distribution was selected for its accuracy in representing commuter patterns. Parameters were calibrated to reflect the sensitivity to distance and destination attractiveness.
Scenario Analysis
Three scenarios were evaluated:
Scenario A: A 50% job reduction in Canary Wharf led to a 36.9% decrease in incoming flows, showing minimal network-wide impact.
Scenario B1 & B2: Increases of 400% and 900% in network-wide transport costs significantly altered commuter patterns, with the 900% increase causing the most substantial flow redistribution.
Analysis
The scenarios highlighted that increased transport costs had a more severe impact on commuter flows than job reductions. This was evident in the more pronounced negative impact on accessibility and attractiveness of the Underground under high cost scenarios.
Conclusion
This project provides a comprehensive analysis of the London Underground's resilience, revealing critical vulnerabilities and the significant effects of both node removals and hypothetical scenarios on network performance. The findings underscore the importance of key stations and the potential disruptions that can arise from changes in job availability and transport costs.
