Tutorial — EV Emission Scenario#

This tutorial varies the electric-vehicle (EV) ratio in the vehicle fleet to quantify how fleet composition affects emissions. It demonstrates the vehicle type editing workflow and the cross-scenario SQL strength of AgentSUMO.

Step 1 — Generate a Seoul baseline#

Simulate traffic in Gangnam Station, Seoul within a 1 km radius for one hour
under medium traffic conditions. Save it as "ev_0pct".

The default vehicle type distribution is conventional combustion vehicles. This serves as the 0% EV baseline.

Step 2 — Create the 25% and 75% scenarios#

For the network and routes of ev_0pct, create two variants:
- ev_25pct: 25% of vehicles are electric
- ev_75pct: 75% of vehicles are electric
Run both and store them under their respective simulation_ids.

The agent will:

  1. Classify as Complex (multi-step, multiple scenarios).

  2. For each ratio, call vehicle_type_edit_tool with the corresponding electric_ratio parameter against the existing route file. This produces a modified route file with the new fleet composition.

  3. Re-run sumo_runner + xml_to_sqlite_tool for each.

Tip

EV scenarios reuse the same network and OD matrix — only the vehicle types change. This is the cheapest kind of variant to produce and the most informative for emission analysis.

Step 3 — Cross-scenario emission comparison#

Compare total CO2, NOx, and PMx emissions across ev_0pct, ev_25pct, and
ev_75pct. Show absolute values and percentage changes versus the baseline.

The agent issues a single SQL query joining the three simulation_ids in the trips table (or alternatively edge_metrics for spatial breakdown), sums the per-vehicle emission columns, and returns a 3×3 comparison table. The emission columns come straight from SUMO’s HBEFA-based emission model — see the schema for trips in Database Schema.

Step 4 — Per-vehicle insight#

For the 75% EV scenario, which vehicle types contributed most to total NOx
emissions? Group by fuel type.

The agent joins trips to vehicle_info on (simulation_id, trip_id = vehicle_id) and groups by vehicle_info.fuel_type. Electric vehicles report zero tailpipe emissions; the remaining NOx comes from the residual combustion fleet.

Step 5 — Visualize#

Show the per-edge CO2 emission heatmap for ev_0pct and ev_75pct side by
side using the difference mode.

The comparative heatmap visualizes the spatial distribution of emission reduction. Edges with steep declines (highways, ring roads) versus edges where reductions are subtle (local streets where idling is dominant) become immediately visible.

What you’ve practiced#

  • Fleet composition editing without re-generating routes.

  • Three-way numeric comparison via SQL on the trips table.

  • Grouping by vehicle_info.fuel_type for fleet-level insight.

  • Spatial difference visualization for emission distributions.