Documentation

VT Planner Analysis Interpretation Guide

Learn how to keep VT Planner scenarios comparable and interpret calculation, simulation, and multi-run evidence.

A useful comparison changes one intended variable while holding the other assumptions constant. Confirm the scenario before reading the headline result.

Match the Method to the Question

Question Use
Is the group broadly sized for the assumed demand? Calculation for RTT, interval, and handling capacity.
How do queues, dispatch, loading, or transfers behave over time? Simulation.
Would the conclusion change with another random realization? Multi-run simulation.

Calculation and Simulation answer related but distinct questions. Do not treat their metrics as interchangeable or expect a single simulated sample to reproduce an analytical value exactly.

Controls That Change the Scenario

Control Why it matters
Demand profile or schedule Changes arrival intensity, direction, origins, destinations, or time periods.
Dispatcher Changes how calls are assigned to cars.
Duration (min) Changes the observation window and whether queue growth is visible.
Seed Changes the random realization while leaving the modeled scenario unchanged.
Capacity factor (%) Changes the usable passenger load of each car.
Load bypass (%) Changes when loaded cars stop accepting additional calls.
Stair factor (%) Changes how much short-trip demand is diverted to stairs before elevator service is simulated.
Settings by car Change the capacity, motion, door, dwell, or transfer behavior of individual cars.
Energy model and inclusion Change the electrical assumptions and whether energy is estimated for the run.
Service zones and transfers Change direct service, passenger paths, and queue locations.

Read Passenger Times Correctly

Result Interpretation
Calculation — Average waiting (AWT) An analytical waiting estimate derived from interval and loading assumptions.
Calculation — Average transit (ATT) An analytical estimate from entering the car to reaching the average destination, including modeled travel, stops, and passenger transfer time.
Calculation — Time to destination (ATTD) AWT + ATT, ending when the destination doors begin opening.
Calculation — Average journey (AJT) ATTD plus the final passenger-transfer allowance.
Simulation — Average Waiting Time (AWT) Time from passenger arrival at each landing until the effective boarding door cycle begins opening, accumulated across transfer legs.
Simulation — Time to Board Time from passenger arrival at each landing until actual boarding, accumulated across transfer legs.
Simulation — Transit Time from actual boarding to alighting, accumulated across journey legs.
Simulation — Journey Time from initial passenger arrival through final alighting, including waits, boarding, travel, and transfer legs.
P90, P95, P99 The value at or below which 90%, 95%, or 99% of the observed passenger times fall.

Calculation estimates and observed Simulation times are related but not interchangeable. Averages describe the center of a result; percentiles reveal the tail. Inspect both.

Check Stability

Use the result as a group of signals:

  • Completion rate: whether generated passengers completed their trips within the result horizon;
  • queue charts: whether waiting demand clears, stabilizes, or grows;
  • passenger activity: whether the tested traffic is the traffic you intended;
  • waiting and journey percentiles: whether a minority of passengers experience very long service;
  • car loads and stops: whether performance relies on heavy loading or excessive stopping;
  • main-terminal departure headway, when available: actual spacing between qualifying departures from the main terminal.

Single Run and Multi-run

Use the same explicit seed for a direct single-run A/B comparison. Use Multi-run simulation to examine variation across samples.

Multi-run repeats the scenario across seeds:

  • the mean summarizes the central tendency across valid runs;
  • pooled distributions and percentiles summarize passenger outcomes across the repeated study;
  • a confidence interval expresses uncertainty in the estimated metric.

A narrow confidence interval indicates low sampling uncertainty under the modeled assumptions. Validate the demand model and equipment assumptions separately against project evidence.

Like-for-like Comparison

  1. Use the same saved building version.
  2. Confirm equivalent demand in each frozen scenario.
  3. Keep duration, dispatcher, capacity factor, load bypass, Stair factor, and zoning fixed unless one is the study variable.
  4. Use equivalent car-specific and energy assumptions, including the same energy-inclusion choice in every compared run.
  5. Use the same seed for single simulations or the same run count and seed policy for Multi-run.
  6. Compare runs with the same completed horizon; use full-duration runs for full-duration comparisons.
  7. Compare the same metric definitions and units.
  8. Read averages, tails, completion, and queues before deciding.

See Results for report actions, charts, replay, exports, and partial-run handling. See Operational Energy for the additional controls needed in an energy comparison.