Communication Materials

Do you have slides, white papers, or other content that has been useful to describe planning under uncertainty? Share them here!

Decisions for the Decade

This game has been presented in several conferences and at individual agencies to great success. Example slides and summaries are below.

The slide deck prepared by the original game creators

The slide deck presented at MoMo 2025 that walks through the game step by step, along with the game board

TRB 2024 Annual Meeting - Future Uncertain Workshop

MoMo 2025 - Uncertainty Workshop

Boston MPO Uncertainty

WFRC Uncertainty

Coming soon…

Software

TMIP-EMAT (Exploratory Modeling and Analysis Tool)

Overview

TMIP-EMAT is an open-source Python framework developed under FHWA’s Travel Model Improvement Program designed to bridge traditional travel forecasting models with DMDU methodologies. Rather than running a heavy regional model for just one or two scenarios, TMIP-EMAT automates experimental design across wide parameter spaces and builds machine-learning metamodels (surrogates). These surrogates rapidly approximate core model outputs, allowing planners to explore thousands of policy levers and uncertain future conditions in seconds.

Key Capabilities

  • Automated Experimental Design: Systematically samples deep uncertainties (e.g., remote work trends, AV adoption, fuel pricing) alongside local policy choices.
  • Metamodel Acceleration: Trains fast-running statistical surrogates on core model runs to eliminate computational bottlenecks during scenario exploration.
  • Scenario Discovery & Policy Optimization: Employs Patient Rule Induction Method (PRIM), CART decision trees, and directed search to uncover the exact conditions under which specific transportation plans succeed or fail.

TMIP-EMAT is not being actively supported by FHWA, but is in use. Responses to a 2024 survey of 14 users is available here.

VisionEval

Overview

VisionEval is an open-source, multi-agency strategic planning framework built in R for rapid scenario evaluation and high-level policy analysis. Structurally designed as a “disaggregate demand / aggregate supply” system, VisionEval models synthetic households and travel behavior in detail without requiring computationally intensive network assignment. This lightweight architecture allows agencies to evaluate hundreds of land-use, pricing, technology, and investment combinations in minutes.

Key Capabilities

  • Rapid Policy Screening: Evaluates regional VMT, greenhouse gas emissions, vehicle fleet turnover, household costs, and equity impacts across wide scenario sets.
  • Behavioral & Demographic Granularity: Captures complex household demographic interactions and behavioral shifts without full-scale activity-based model runtime overhead.
  • Ideal DMDU Testbed: Serves as a fast-running standalone strategic tool or as a core model within TMIP-EMAT to filter scenario spaces before running full regional models.

Get Involved

Zephyr is a community effort — join as a member, support our activities, or say hello in the Zephyr Forum.

Become a Member Donate Join the Zephyr Forum