How we can reduce traffic congestion
Google is testing changes to its Maps routing algorithm that deliberately steer some drivers off chronically congested road segments, yielding modest gains in average speeds and fuel use but also raising questions about fairness, consent, and trust in navigation apps. Commenters note that dynamic routing can unintentionally flood residential streets, overuse roads not designed for heavy traffic, and create perverse “Google detours,” while others point out parallels to concepts like Braess’s paradox and induced demand. Many argue that algorithmic tweaks can only marginally ease congestion, emphasizing instead structural solutions such as better public transit, walkable land use, congestion pricing, and expanded remote work.
Overview of Google’s Congestion-Dispersion Experiment
- Modified routing to steer drivers off pre-selected congested segments toward similar-time alternatives.
- Tested via city-wide alternating “treatment/control days” over six months in 10 major US cities.
- Reported median ~2% speed increase on targeted segments and ~0.5–1% reduction in fuel consumption.
- Some commenters find the impact positive but modest; others say it’s surprisingly late and obvious “load balancing.”
Routing Logic, Game Theory, and Data Collection
- Braess’s paradox and “everyone optimizes for themselves” are cited as reasons explicit load balancing can beat naïve “fastest route” choices.
- Some argue congested segments already become unattractive in routing; others note algorithms still favor major roads that are slightly faster “on paper.”
- Several suspect apps sometimes route users down suboptimal or little-used roads to refresh traffic data; others say if it happens, it’s only when alternatives are marginal.
Ethics, Oversight, and City Coordination
- Debate on whether rerouting millions of people counts as human-subjects research needing IRB-style oversight.
- Some argue big routing changes with community-level externalities should involve city planners; others warn against overestimating benefits of government review.
- Concerns about trucks being diverted onto weaker or residential roads, increasing road damage and local backlash.
Trust, User Experience, and “Greater Good” vs Individual Optimum
- Many report “Google detours” that save 2–4 minutes in theory but in practice cause longer, unsafe, or bizarre routes, sometimes through alleys or across difficult turns.
- Perception that recent routing has worsened leads some to abandon navigation for familiar trips or switch apps.
- Several object philosophically to being sent on slower routes for collective benefit without explicit consent.
- Requests for more user controls (e.g., “avoid left turns”) and less intrusive mid-trip rerouting prompts.
Side Effects on Neighborhoods and Infrastructure
- Complaints that navigation apps already turn quiet residential streets into “rat runs,” blocking driveways and impeding emergency vehicles.
- Worries that diffusion-like strategies could further overload streets not designed for commuter or heavy-truck traffic.
- Some suggest road design, zoning, and pricing should handle such flows, not opaque corporate algorithms.
Alternative or Complementary Solutions
- Strong advocacy for:
- More and better public transit, trams/metros, and protected bike infrastructure.
- Congestion pricing, with revenue used to improve and possibly make transit free.
- Smarter traffic signals and better lane design (including “road diets”).
- Remote work to reduce commuting demand.
- Counterpoints emphasize cost overruns, induced demand (or “suppressed demand”), equity concerns, and differing lifestyle preferences (suburban vs dense, walkable areas).