Zone Based Structured Scheduling: A Case Study in Lee County, North Carolina

Transcription

1 Zone Based Structured Scheduling: A Case Study in Lee County, North Carolina 206 International Conference on Demand Responsive Transportation Jeremy Scott ITRE Public Transportation Group jscott@ncsu.edu

2 Demand Response Transportation Typical Demand Response Transportation Inefficient to deliver Driver schedules vary every day Difficulty understanding why service and costs vary Time-consuming to schedule Bigger emphasis on data driven analytics As demand response services increase, transit systems experience capacity constraints for: Funding Scheduling Vehicles Drivers

3 Solution? Efficiency Scheduler All in one Solution Drivers Structured Scheduling Simplifies scheduling and dispatching process Improves efficiency Reduces cost per trip Increase ridership by offering organized and understandable service at established times Customers Agency

4 Ideal Candidate? Client Home Client Home Client Home Available Services Client Home Client Home

5 Lee County and the City of Sanford Sanford Population: 29,470 Lee County Population: 60,266 90% of destinations in Sanford 89% of all trips have destination in Sanford

6 How are Initial Zone Boundaries Created?. Start with county boundary 2. Overlay roads 3. Overlay waterbodies 4. Overlay railroads 5. Rasterize all layers 6. Determine central & boundary points 7. Assign weight to each layer, sum, & calculate least cost path 8. Create central zone 9. Combine central zone with initial boundaries 0. Overlay customers. Finalize zone boundaries

7 Final Zone Boundaries Institute for Transportation Research and Education

8 Scheduling Part Trips Begin & End in Central Zone 55% Service Time Assign Vehicles Time S Time S Time S Time S S2 S3 S4 S

9 Scheduling Part 2 Trips Between Outer Zone & Central Zone 4% Service Time Assign Vehicles Time S-->B B-->S B-->B Time S-->B B-->S B-->B Time S-->B B-->S B-->B Time S-->B B-->s B-->B 4 5 x x x x

10 Scheduling Part 3 Trips Between Outer Zones 4% Time W-->P P-->W W-->T T-->W C-->B B-->C B-->L B-->F D-->B Service Time Sweeper Assign Vehicles Sweeper Sweeper 2

11 Complete Schedule Time S S2 S3 S4 S5 S->W W-->S S-->P P-->S S-->T T-->S S-->L L-->S S-->C C-->S S-->B B-->S S-->F F-->S S-->D D-->S Sweepers

12 Driver Schedule A-A A-C A-F Sweeper A-B A-D A-G Available A-B A-E A-H Driver Total Drivers

13 Final Zones with Service Times

14 Steps to Implementation Approval from County Commissioners and/or Transit Advisory Board (TAB) Driver/Scheduler Training Sessions Very important for driver buy-in Start testing new scheduling process to minimize issues at go-live Funding agency outreach

15 Steps to Implementation Cont d Branding of Service: MUSTANGS Public Outreach Media blitz Newspaper advertisements Radio Interviews Attendance at community events Creation of Index cards

16 Results Compare pre-implementation efficiency to postimplementation efficiency Trips Trips/Hour

17 Historical Trips per Day

18 Data Sets Used for Comparison Implementation started on January 4 th, 206 Eliminated January due to growing pains February due to adverse weather Used March July for pre- and post-implementation periods

19 Summary Statistics Pre-Implementation Post-Implementation st Quartile: Trips/Hour rd Quartile: Trips/Hour IQR: Trips/Hour Upper Outlier Lower Outlier.60.5 Average Trips/Hour with outliers Average Trips/Hour without outliers Average Trips/Day Average Trips/Day Range [64,264] [2,234]

20 Trips vs. Trips per Hour Pre-Implementation Post-Implementation 3//5 7/3/5 3//6 7/3/ y = x y = x Average Trips per Day: 28 Average Trips per Hour:.98 Average Trips per Day: 82 Average Trips per Hour:.90

21 Trips vs Trips per Hour Institute for Transportation Research and Education Pre-Implementation Post-Implementation Linear (Pre-Implementation) Linear (Post-Implementation) y = x y = x

22 Trips vs Trips per Hour: Scaled to Same Trips Range Pre-Implementation Post-Implementation Linear (Pre-Implementation) Linear (Post-Implementation) y = x y = x

23 Conclusions Structured Scheduling more efficient at lower trips/day As trips increase, Structured Scheduling becomes less efficient relative to old method Structure based on higher trip count Trips lost to private transportation provider Drivers not following manifests Customers enjoy more understandable scheduling environment Willing to schedule appointments based on schedule Structure allowed COLTS to find open times to add new service Worked with Hispanic community (BIC Grant)

24 Re-evaluate current schedule Next Steps? Overlay existing demand to see if changes in structure have occurred Look for seasonality

25 Next Steps? Institute for Transportation Research and Education