INTERACTIVE EXPERIMENT / RULES & ASSUMPTIONS
The transport policy lab
Try a different rule. Then ask how much of the result comes from the model’s assumptions.
Teaching simulation · not a forecastSTRUCTURE / WEEK 0
UNDER THE SURFACE
The rules are part
of the result.
Change the settings, inspect the trajectory, and export the values to examine them yourself.
Model & assumptions
Transit probability = 1 / (1 + exp[−sensitivity × (transit utility − car utility)])An illustrative closed population of 100,000 commuters starts with 25% using transit. Each 0.25-week step moves a fraction toward a logit choice probability. Car utility = 2.1 − 1.4 × (cars / road capacity)² − 0.35 × pricing − 0.25 × information. Transit utility = 0.6 + quality + 0.45 × fare reduction. Baseline road capacity is 60,000. Quality starts at 0.8 and changes with delayed investment minus 0.04 × quality per week. Base investment is 0.035/week; pricing adds 0.04 + 0.025 × car share. These coefficients are invented teaching assumptions. Prices, budgets, quality and utility are normalized indices. Only mode switching is represented: induced trips, land use, equity, route coverage and construction costs are omitted. Road widening therefore cannot establish real-world induced-demand magnitudes. The baseline uses the same behavioral settings with all interventions off.