Skip to content

Tensions in Practice: A shortcut can slow flow and preserve a backup

A four-node route network

Imagine one unit of flow from S to T through intermediate nodes U and V. S→U and V→T each cost their own edge load x; U→T and S→V each cost 1. An optional U→V connector costs 0. Travelers are infinitesimally small and independently choose cheapest routes, with full information. Compare closing the connector with leaving it open. Then inspect a separate failure scenario in which both constant-cost edges disappear. The connector changes both the ordinary routing equilibrium and which paths survive that failure.

Reduce ordinary routing cost

Remove the option that draws self-routed flow onto both load-sensitive edges.

Retain the failure path

Keep the connector so a route remains if both constant-cost edges fail.

Why these aims pull against each other

The connector is harmful to the modeled ordinary equilibrium but useful in the specified failure state. An equilibrium comparison alone cannot price that reserve.

Compare the arrangements

Close the connector

Normal flow splits equally between S–U–T and S–V–T. Each route costs 0.5 + 1 = 1.5. Removing both constant-cost edges leaves no complete path.

Closed: cheaper equilibrium, no failure path.
Normal flowRoute costAfter failures
S–U–T0.51.5Unavailable
S–V–T0.51.5Unavailable
S–U–V–TNot a routeNot a routeUnavailable
What it protects
All normal flow experiences cost 1.5 rather than 2 under the stipulated self-routing model.
What it costs
The specified simultaneous failure disconnects S from T.
When it fits
Plausible if lower ordinary cost warrants accepting this particular loss of backup reachability or another failure plan exists.

Illustration note: This is an editorial, deliberately bounded illustration. Its stated rules and any numbers are invented, not observations, recommended settings, or predictions.

Keep the connector

Normal flow uses S–U–V–T, putting load 1 on both load-sensitive edges and giving cost 2. Either unused side route also costs 2, so no individual has a cheaper alternative. After the two constant-cost edges fail, the connector route still exists.

Open: costlier equilibrium, surviving route.
Normal flowRoute costAfter failures
S–U–T02Unavailable
S–V–T02Unavailable
S–U–V–T12Available
What it protects
A path survives the explicitly named simultaneous failure.
What it costs
Normal equilibrium cost rises from 1.5 to 2; keeping an extra route does not force self-routing to use it well.
When it fits
Plausible if this failure path is sufficiently valuable to justify the normal-time penalty under the stated routing rule.

Illustration note: This is an editorial, deliberately bounded illustration. Its stated rules and any numbers are invented, not observations, recommended settings, or predictions.

What this illustration does—and does not—establish

The source establishes the structural tension; the concrete alternatives and their conditional costs are editorial synthesis. No arrangement is a universal recommendation.

  • This is an invented nonatomic equilibrium model. Costs are arbitrary units; no travel-time observations, failure probabilities or policy recommendation are supplied.
  • Normal route costs are evaluated at each arrangement’s displayed normal flow. The failure column shows reachability only, not an additional equilibrium calculation.
  • A coordinated dispatcher could leave the open connector unused and retain the cheaper split. Convergence, implementation cost and mixed routing behavior are outside this comparison.

Source entries

Braess's Paradox

Prime · Source of the tension

Braess paradox Addition versus removal symmetry (sign/direction) supplies the local tension. The setting, alternative arrangements, and stipulated consequences are editorial applications.

Addition versus removal symmetry (sign/direction)

Diagnostic: before cutting, ask what the element does off-equilibrium under failure or surge; an edge useless at equilibrium may be essential under disturbance.

Read the source section

Self-routing is a load-bearing assumption

The paradox lives entirely in the gap between the decentralized equilibrium and the achievable optimum; it presupposes routing is selfish.

Read the source section