Transit Stop Consolidation Evaluation of State of Practice

ADEC GIS-T 2011, St. Petersburg, Fl 1

Transit Stop Consolidation Evaluation of State of Practice

GIS in Transit 2011St. Petersburg, Florida

Sep. 13-15, 2011Maaza Mekuria, PhD, PE, PTOE

ADECorpSan Jose, CA

ADEC GIS-T 2011, St. Petersburg, Fl 2

Modeling Paradigms

• Continuous – Pros

• Useful for exploring relationships between variables (demand, spacing, volume)

– Cons • Lacks sensitivity to

– Location/safety– demand concentration– Access network geometry

ADEC GIS-T 2011, St. Petersburg, Fl 3

Modeling Paradigms

• Discrete – Pros (can be made to be sensitive to)

• Physical location• Demand intensity• Access network geometry

– Cons • Requires

– More detailed data – Robust analytical tools– More time

ADEC GIS-T 2011, St. Petersburg, Fl 4

Stop Spacing Impacts

optimum

0

5

10

15

20

25

30

35

200 240 280 320 360 400 440 480 520 560 600 640

Stop Spacing (m)

Tim

e (m

in)

Access Time

In-Vehicle Time

Total Impact

Oper. Time

Region "A" Region "C" Region "B"

ADEC GIS-T 2011, St. Petersburg, Fl 5

Modeling Stop Spacing

• Single/Bi-Directional Stops

• Multiple Periods

• Sensitivity to Various Input Parameters

• Common Base for comparison

• What if scenarios

ADEC GIS-T 2011, St. Petersburg, Fl 6

State of the Practice

• Rule of Thumb based on– Demand at a stop– Proximity to the nearest stop– Other policy considerations

ADEC GIS-T 2011, St. Petersburg, Fl 7

Stop Spacing Optimization

• Geographic modeling – Walk impact using real street network– Demand density using parcel attributes– Stop enumeration– Multi-period analysis – Sensitivity analysis

ADEC GIS-T 2011, St. Petersburg, Fl 8

Service Area Analysis using three data sets

ADEC GIS-T 2011, St. Petersburg, Fl 9

DP Immediate Cost Computation

• To find costs at j, need to specify at least 1 or more predecessor successor stops

• >=1-dimensional search space

ADEC GIS-T 2011, St. Petersburg, Fl 10

Quintuplet DP Algorithm

)},,;(

),,,;(),,,;(),,,;({),,;( min)(max)(min

mkjlf

mljikopermljikridemljikwalkljikfkSmkS

walk(k; i,j,l,m) - sum of walking cost impacts associated with stop k over all periodsride(k; i,j,l,m) - sum of riding cost impacts associated with stop k over all periodsOper(k; i,j,l,m) - sum of operating cost impacts associated with stop k over all periodsf(l; j,k,m) – optimal return function cost impacts associated with stop l to the end.

ADEC GIS-T 2011, St. Petersburg, Fl 11

Study Area I Transit Details • Route 57 Crosstown Bus

– Kenmore to Watertown 45 Inbound and 43 Outbound existing stops• Peak ridership

– AM East Bound to Kenmore– PM peak Westbound to Watertown

ADEC GIS-T 2011, St. Petersburg, Fl 12

Study Area II Transit Details • Route 1

Crosstown Bus

35 SB, 36 NB existing stops

• Peak ridership

– Balanced in both directions

ADEC GIS-T 2011, St. Petersburg, Fl 13

Rt 57 Cross-town Bus DP Result Comparison Vs Consultant

Recommendation (Inbound dirn.)

number of stops (Inbound)

Change in Walk Time (Pax-Min/Day)

Change in Running Time (min)

Change in Walk Cost ($/Day)

Change in Ride Cost ($/Day)

Change in Operating Cost ($/Day)

Change in Total Cost ($/Day)

Existing 45

Analyst recommendation 31 7669 -1.9 $3,672 -$2,509 -$488 $674

DP Optimum 32 3955 -1.8 $1,761 -$2,211 -$469 -$920

ADEC GIS-T 2011, St. Petersburg, Fl 14

Rt 1 Cross-town Bus DP Result Comparison Vs Consultant

Recommendation

Total 2-way stops

Average walk time (min)

Average ride time (min)

Avg two-way running time (min)

Change in Walk Cost ($/day)

Change in Ride Cost ($/Day)

Change in Oper Cost ($/Day)

Change in Total Cost ($/Day)

Existing 71 3.46 36.4 88.5 0 0 0 0

Analyst recommendation 55 3.77 35.7 87.4 705 -267 -113 326

DP Optimal 45 3.57 35.5 86.1 243 -218 -156 -132

ADEC GIS-T 2011, St. Petersburg, Fl 15

Summary

State of the Practice can not capture all factors

Use of standard data sets

Appropriate tools with flexibility

Analyst input

Geographic Output

ADEC GIS-T 2011, St. Petersburg, Fl 16

Questions?