I've watched infrastructure planners make the same mistake for years. They optimise a single corridor, reduce its cost, improve its efficiency, and then watch the entire network perform worse than before.

This isn't incompetence. It's a structural misunderstanding of how networks actually behave.

The assumption is simple: if you improve individual components, the system improves. But networks don't work that way. They operate through interdependencies, feedback loops, and emergent behaviours that single-corridor optimisation can't capture. When you focus exclusively on one route, you're solving a problem that doesn't exist in isolation.

The result is predictable. You add capacity where it seems logical, only to discover that adding infrastructure has increased congestion, reduced throughput, and created bottlenecks elsewhere in the system.

This is the corridor illusion. The belief that infrastructure planning is about optimising individual routes when it's actually about understanding how routes interact across the network.

The Paradox That Breaks Conventional Planning

Dietrich Braess demonstrated this in 1968 with a traffic network problem that still confounds infrastructure planners today. He proved mathematically that adding a new road to a congested network can increase travel times for everyone.

The mechanism is straightforward. When you introduce a new route, individual users make selfish routing decisions that optimise their own journey times. Each person chooses the path that appears fastest based on current conditions. But as more users make the same calculation, the new route becomes congested, and the collective outcome is worse than the original network configuration.

This isn't theoretical. Steady increases in natural gas transportation volumes have prompted operators to re-evaluate existing pipeline infrastructure, where the paradox challenges the traditional understanding that additional links enhance capacity.

The same principle applies to electricity transmission, water distribution, and hydrogen networks. Adding capacity in one location shifts flows elsewhere, creating cascading effects that undermine the original optimisation.

The problem isn't the new infrastructure. The problem is the assumption that local improvements translate into system-wide benefits.

Efficiency and Resilience Are Structurally Incompatible

When you optimise a network for efficiency, you reduce its resilience. When you optimise for resilience, you sacrifice efficiency. This isn't a design choice you can engineer around. It's a fundamental constraint of network topology.

Research shows that very efficient networks are not resilient whilst very resilient networks lack efficiency. Best efficiency is realised in star-like configurations, whilst enhanced resilience is related to the avoidance of short loops and degree homogeneity.

This creates a structural tension. Star configurations minimise total path length and reduce capital expenditure, which is why transmission owners favour them. But they also create single points of failure. When a central node fails, the entire network collapses.

Resilient networks require redundancy. Multiple pathways between nodes. Distributed connectivity. Higher capital costs. Longer average route lengths. All of which reduce efficiency.

The trade-off is measurable. A redundant network gives a higher probability of flow between any two randomly chosen nodes, whereas an efficient network restricts and constrains flow choices. The ratio between system redundancy and efficiency determines resilience in complex infrastructure systems.

You can't optimise for both simultaneously. You have to choose which constraint matters more under specific operating conditions.

Cascading Failures Expose the Limits of Component-Level Thinking

Infrastructure doesn't fail in isolation. When one asset becomes disabled, failures propagate to dependent assets, creating cascades that lead to system collapse.

This is where corridor-level optimisation becomes dangerous. You design a route to handle expected loads under normal conditions. You model demand, calculate capacity, and ensure the corridor meets technical specifications. But you don't model what happens when adjacent infrastructure fails and demand shifts onto your optimised route.

The result is predictable. A failure in one part of the network creates flow redistribution. Your optimised corridor, which was designed for steady-state conditions, now experiences loads it wasn't built to handle. It fails. The cascade continues.

A resilience-based design framework represents a shift from traditional load-based approaches to a systems-thinking methodology that emphasises the critical importance of assessing cascading impacts and regional interdependencies before determining appropriate mitigation strategies.

This requires modelling the network as an interdependent system, not as a collection of independent corridors. You have to simulate failure scenarios, test redistribution patterns, and identify which routes become critical under stress conditions.

Corridor optimisation can't capture this. It operates at the wrong scale.

Selfish Routing Creates Collective Inefficiency

When users decide selfishly upon their route through the network, the system reaches a stable state known as Nash equilibrium. At this point, travel times are equal for all individuals, and any change of route would increase their travel times.

This is the user optimum. It's stable, predictable, and suboptimal.

The system optimum minimises the maximum travel time in the network and often leads to lower travel times overall. But it requires coordination. Users must accept individually suboptimal routes to achieve collective efficiency.

Infrastructure networks face the same problem. Transmission owners, distribution operators, and independent generators make routing decisions based on their own constraints. Each decision is locally rational. But the aggregate outcome is a network configuration that's stable yet inefficient.

You can see this in power grids. The post-disaster network should contain as many available components as possible to give the best topological connection. But research shows this assumption can be counterproductive. More infrastructure doesn't automatically equal greater resilience.

The issue is coordination. Without system-level oversight, individual optimisation behaviour creates collective inefficiency. This isn't a failure of individual decision-making. It's an emergent property of decentralised routing in complex networks.

Global Trade Networks Validate the Efficiency-Resilience Tension

The efficiency-resilience trade-off isn't confined to energy infrastructure. It shows up in global food trade networks, supply chains, and communication systems.

Increased connectivity of global food trade increases its efficiency at the expense of its resilience to spreading risk. Networks prioritising efficiency become more vulnerable to targeted node attacks during periods of sparsity.

This empirical finding validates that the tension isn't theoretical. It's a measurable, real-world constraint affecting critical infrastructure worldwide.

When you optimise trade routes for cost and speed, you concentrate flows through a small number of high-capacity corridors. This reduces redundancy. A disruption in one corridor has cascading effects across the entire network because alternative pathways don't exist or lack sufficient capacity.

The same pattern emerges in electricity transmission. Optimising individual corridors for least-cost routing creates networks with concentrated flows and limited pathway diversity. When demand spikes or infrastructure fails, the network lacks the flexibility to redistribute loads effectively.

This is the corridor illusion in practice. You solve the local problem whilst creating a system-level vulnerability.

What Systems-Level Planning Actually Requires

You can't plan infrastructure by optimising individual routes. You have to model the network as an interdependent system where local decisions create emergent behaviours.

This requires three shifts in methodology.

First, you model pathway diversity, not individual route efficiency. The question isn't whether a single corridor is optimal. The question is whether the network maintains sufficient redundancy to absorb disruptions without cascading failures.

Second, you simulate failure scenarios and test redistribution patterns. You identify which routes become critical under stress conditions and ensure the network can handle load shifts without collapse.

Third, you quantify the efficiency-resilience trade-off and make explicit decisions about which constraint matters more under specific operating conditions. You don't pretend you can optimise for both simultaneously.

This is harder than corridor-level planning. It requires computational methods that can model network-scale interactions, evaluate multiple scenarios simultaneously, and incorporate non-technical constraints like stakeholder opposition and consenting risk.

But it's the only approach that reflects how infrastructure networks actually behave. Corridor optimisation produces locally rational decisions that create systemically irrational outcomes. Systems-level planning produces decisions that account for interdependencies, feedback loops, and emergent behaviours.

The corridor illusion persists because the tools haven't caught up to the problem. Most infrastructure planning software operates at project scale, not network scale. It optimises individual routes without modelling how those routes interact across the system.

That's changing. Computational methods now exist that can model pathway diversity, simulate cascading failures, and quantify the efficiency-resilience trade-off in real time. The question is whether infrastructure planners are ready to use them.

Move Beyond the Corridor

Stop optimising individual routes. Start modelling the network as an interdependent system where local decisions create emergent behaviours.

Test your infrastructure plans against failure scenarios. Simulate load redistribution. Identify which routes become critical under stress conditions.

Quantify the efficiency-resilience trade-off and make explicit decisions about which constraint matters more under your operating conditions.

The corridor illusion breaks when you recognise that network value comes from pathway diversity, not individual route efficiency.

‍