Integrated Travel Demand Model Challenges and Successes Tim Padgett, P.E., Kimley-Horn Scott Thomson, P.E., KYTC Saleem Salameh, Ph.D., P.E., KYOVA IPC.

Integrated Travel Demand Model

Challenges and Successes

Tim Padgett, P.E., Kimley-HornScott Thomson, P.E., KYTC

Saleem Salameh, Ph.D., P.E., KYOVA IPC

2015 WV Planning ConferenceSeptember 16

Davis, West Virginia

“There is really no hope that a mathematical model can ever accurately predict the future, given the uncertainty in demographics, technological shifts, and social changes.”

– JD Hunt

“But we still try.” – Tim

Forecasting Background

• Why do we use forecasting models?

Model Background

• History

• Model versions over time

• Previous model and transition to current

2015 KYOVA Travel Demand Model

Useful Planning Tool • CMP, MTP/TIP, corridor studies, and other

applications

Three Current Models• KYOVA, Ashland, and RIC

Limitations• Not all are time-of-day

• Speeds and capacities calculated differently

• “Apples vs. oranges”

Model Update• Expand KYOVA model to include Ashland model

• Integrate HCM 2010 methods for speed, capacity

• Incorporate HERE travel time data

• Update external travel data

• Incorporate truck network, trip matrix

• V/C calculations

• Documentation and training

Additional Features

• Free-flow speed override

• Simplified user interface

Automatic Reporting

• RMSE

• VMT by county and functional class

• Congested speeds by county and functional class

• VHT by county and functional class

Travel Demand Model Integration Issues

• Network attributes- Integrate state linear referencing systems (Road analyzer, HIS,

and LRS)

• Household model- Household composition and trip rates for additional counties

• Employment data- Coordinate employment classes and data sources

• Traffic counts- 2010 daily, hourly, and truck counts for calibration and validation

• Trip distribution- Re-estimate gravity model parameters to reflect expanded area

Challenges and Successes

Challenges• Data

• Multiple jurisdictions

Successes• Coordination and cooperation

Thinking Beyond the Model

When the Model isn’t Enough

• Travel demand models don’t do a great job with:- Transit and Transit Suitability

- Freight Planning

- Alternatives Analysis

• Balance between complexity of model, ease of use, and cost

Mode Share

What is it?

• District of Columbia Multimodal Long Range Transportation Plan

What was the question?

• How much can we increase non-auto mode share by expanding transportation choices and improving the reliability of all transportation modes?

Mode Share

How did we answer the question?

• Used elasticities to account for introduction of new service and expansion of existing service

• Used GIS to determine influence areas and applied this to our model trip tables to shift trips between modes

Mode Share

Mode Share

Results and Conclusions

• Answers would not have come from the travel demand model alone

• Interesting—showed that investment alone wasn’t enough to achieve their goal of 75% non-auto mode share

KYOVA’s Spatial Decision Support System (SDSS)

Integration

What is a Spatial Decision Support System (SDSS)?

“…an interactive, flexible,

and adaptable computer-

based information system,

especially developed for

supporting the solution of

a non-structured

management problem for

improved decision

making. It utilizes data,

provides an easy-to-use

interface, and allows for

the decision maker’s own

insights…” (Turban, 1995)

Spatial Decision Support System Characteristics

Current SDSS Maps

CMP Map Figures• KYOVA Transportation Management Area

boundary

• CMP network

• Major river crossings

• Fixed-route transit coverage

• Computed crash rates compared to statewide average

• Congested locations from stakeholder workshops

• Downtown railroad underpass/viaduct locations

Current DSS Maps

CMP Map Figure Time of Day Years

Volumes from Traffic Model AssignmentsAM Peak, PM Peak, Average Daily

Base Year, 2020, 2030, 2040

Capacities from Traffic Model AssignmentsAM Peak, PM Peak, Average Daily

Base Year, 2020, 2030, 2040

V/C Ratios from Traffic Model AssignmentsAM Peak, PM Peak, Average Daily

Base Year, 2020, 2030, 2040

Levels of ServiceAM Peak, PM Peak, Average Daily

Base Year, 2020, 2030, 2040

Travel Time IndicesAM Peak, PM Peak, Average Daily

Base Year, 2020, 2030, 2040

Planning Time IndicesAM Peak, PM Peak, Average Daily

Base Year, 2020, 2030, 2040

Third Party Data Sources

Traffic Conditions

Trip RoutesFreight Speeds

NPMRDS

National Performance Measure Research Data Set - 6th Region

West Virginia’s Network

November 2014

Link Level Speed Data

Blue = minimum speeds Red = maximum speeds Arrow = direction of travel

Origin-Destination Data

Trips by Day Part

Tim Padgett, P.E., [email protected]

Scott Thomson, P.E., [email protected]

Saleem Salameh, Ph.D., P.E., KYOVA [email protected]

Questions?