(Converted from Microsoft PowerPoint Presentation, approximate slide layout retained)

Large Scale Simulation Modeling of Events for Phoenix

Mark Hickman

ITE National Conference on Transportation for Planned and Unplanned Events

March 2009

Outline

■ Overview of simulation-based platform

■ Scenarios

□ No-notice evacuation during a Cardinal game

□ Short-notice evacuation due to flooding

■ Modeling and Results for flooding scenario

■ Conclusions

Simulation-based platform

■ Congestion is not linear

■ Complex interactions between

□ Evacuees making various decisions

■ Where to go, when to leave, which route to take

■ How to respond to radio or message sign information

□ Roadways/intersections capacity

■ Capacity

□ Traffic management strategies

■ Diversion, reversible/contra flow lanes, information, etc.

■ Simulation allows for systematic and scientific evaluation of these intersections

Simulation Architecture

Diagram showing Supply (Traffic Flow, Network Configurations, Traffic Controls) with Demand (Participation, Departure Time, Route) via Interactions leading to Performance Measure

Flooding Scenario - Short-notice Evacuation

□ Spatial and temporal characteristics

■ Assume dam burst occurs at 12 PM

■ Worse case assumed - five flooding contour map with 5 = highest water level

□ Inundation follows the inundation map from Dept. of Emergency and Military Affairs

■ Flooding warning is broadcasted to citizens at 12:30 PM

■ Evacuees leave as early as 1 PM.

Contributions

Method to estimate demand from existing planning tools

■ Trip matrices by time of day

■ Trip matrices by trip purpose

■ Common simulation platform

Models

■ Trip generation

■ Trip distribution

■ Departure time choice

Demand Modeling Methods

■ Use of existing planning-level travel model information

□ Trip tables by time of day, trip purpose

■ Trip Generation

□ Total number of vehicles generated from the flooding zones

■ Trip Distribution

□ Number of vehicles from flooding zones to non-flooding area

□ Destinations: home, relatives or friends, shelters

□ Home-based other (HBO) trip pattern

■ Departure Times

□ Starting at 12:30 after dam break at 12:00

□ Evacuation time: 1.5 hrs to 5 hrs for each contour

Contour 1: Flooding within 2 hrs

Map showing Affected Segments in Contour (in green lines) near Phoenix

Contour 2: Flooding within 2.5 hrs

Map showing affected segments in Contour 2 (in green lines)

Contour 3: Flooding within 3.5 hrs

Map showing affected segments on Contour 3 (in green lines)

Contour 4: Flooding within 4.5 hrs

Map showing affected segments in Contour 4 (in green lines)

Contour 5: Flooding within 5.5 hrs

Map showing affected segments on Contour 5 (in green lines)

Trip Generation from Flooding Zones

Map graph showing regions of Phoenix based on volume of cars

Trip Generation / Distribution

■ 1,275 North zones, 720 South zones

Trip Distribution to Destinations

Map of Phoenix showing volume of vehicles

Time Frame of Evacuation

Timeline showing evacuation based on Contours 1-5, for flooding

Cumulative Departure Rate

□ Departure Curve

■ All contours will depart 1 hr earlier before flooding.

■ Contour 1 and 2 will depart abruptly and Contour 3, 4, and 5 will depart moderately

Graph showing timeline of contour 1 flooded by minutes

Total Departure Rate

Around 50% of evacuees depart in first 30 minutes

Scenarios

■ Baseline scenario

□ Evacuees receive updated traffic information every 15 minutes, over the radio

□ Background demand to return to their origins in the case of no available route to the destinations

□ The evacuee demand may be separated into north- and south-bound by the flood area

■ Specific management strategies

□ Information update

□ Contra-flow facilities

Behaviors around flooding area

Diagram showing Evacuee demand, Flooding and Destination

Scenarios with Management Strategies

■ "Info" strategy

□ Frequent radio broadcasts of updated traffic information -every 5 minutes

■ "Contra" strategy

□ Contra-flow on three major arterials for movements south from the flood area, in addition to the 5-minute radio updates

Results: Travel Times

Table showing Total Vehicles for all demand and for evacuee only

■ Total and average travel time is decreased with strategies

■ Contra-flow techniques are not as effective - poor downstream distribution

Results: Percent Reaching Safety

Table showing Strategy by each Contour 1-5, on baseline, info and contra

■ Evacuees from early contours (1, 2) have trouble reaching safety, due to network congestion

□ Staging evacuation may be more effective

■ Some improvement for information and contra-flow strategies

Conclusions

■ Developed mesoscopic simulation model

■ Developed demand model with existing planning data

■ Valuable tool for evacuation planning and operations

■ New strategies evacuated vehicles more quickly

■ Simulation platform to be enhanced and other strategies to be studied in Phase II effort

■ Simulate pedestrian traffic at points

■ Route-choice behavior to travel advisories, and

■ Adaptive wide-area traffic management



	Outline ■ Overview of simulation-based platform ■ Scenarios □ No-notice evacuation during a Cardinal game □ Short-notice evacuation due to flooding ■ Modeling and Results for flooding scenario ■ Conclusions



	Simulation-based platform

	■ Congestion is not linear ■ Complex interactions between □ Evacuees making various decisions ■ Where to go, when to leave, which route to take ■ How to respond to radio or message sign information □ Roadways/intersections capacity ■ Capacity □ Traffic management strategies ■ Diversion, reversible/contra flow lanes, information, etc. ■ Simulation allows for systematic and scientific evaluation of these intersections



	Flooding Scenario - Short-notice Evacuation

	□ Spatial and temporal characteristics ■ Assume dam burst occurs at 12 PM ■ Worse case assumed - five flooding contour map with 5 = highest water level □ Inundation follows the inundation map from Dept. of Emergency and Military Affairs ■ Flooding warning is broadcasted to citizens at 12:30 PM ■ Evacuees leave as early as 1 PM.



	Contributions Method to estimate demand from existing planning tools ■ Trip matrices by time of day ■ Trip matrices by trip purpose ■ Common simulation platform Models ■ Trip generation ■ Trip distribution ■ Departure time choice



	Demand Modeling Methods ■ Use of existing planning-level travel model information □ Trip tables by time of day, trip purpose ■ Trip Generation □ Total number of vehicles generated from the flooding zones ■ Trip Distribution □ Number of vehicles from flooding zones to non-flooding area □ Destinations: home, relatives or friends, shelters □ Home-based other (HBO) trip pattern ■ Departure Times □ Starting at 12:30 after dam break at 12:00 □ Evacuation time: 1.5 hrs to 5 hrs for each contour



	Cumulative Departure Rate

	□ Departure Curve ■ All contours will depart 1 hr earlier before flooding. ■ Contour 1 and 2 will depart abruptly and Contour 3, 4, and 5 will depart moderately



	Total Departure Rate Around 50% of evacuees depart in first 30 minutes



	Scenarios ■ Baseline scenario □ Evacuees receive updated traffic information every 15 minutes, over the radio □ Background demand to return to their origins in the case of no available route to the destinations □ The evacuee demand may be separated into north- and south-bound by the flood area ■ Specific management strategies □ Information update □ Contra-flow facilities



	Conclusions

	■ Developed mesoscopic simulation model ■ Developed demand model with existing planning data ■ Valuable tool for evacuation planning and operations ■ New strategies evacuated vehicles more quickly ■ Simulation platform to be enhanced and other strategies to be studied in Phase II effort ■ Simulate pedestrian traffic at points ■ Route-choice behavior to travel advisories, and ■ Adaptive wide-area traffic management



	Thanks!