1. Disabled Adult Transit Service:
Shift Design Optimization
and
Demand Forecasting
Team:
Mike Clay Yvan Fortier Chris Samuel

2. Agenda PART A: Shift Design Optimization Problem Solution Methodology
Analysis
Results
Risk
PART B:
Demand Forecasting
Problem
Solution
Methodology
Analysis
Risk
Results
Problem – Solution – Methodology – Analysis – Results - Risks

3. Executive Summary

4. Executive Summary: Part A
Part-Time
Shifts
Full-Time
Shifts
Daily Cost
and
An increase in the amount of part time shifts and less full time shifts will reduce overtime hours paid out. It will be effective in reducing daily operator cost while still meeting same level of demand
Productivity

5. Executive Summary: Part B
1. Redirecting
2. Optimizing
3. Centralizing
Cost Containment
Strategies:
Operating
Expenses:
197%
Number of
Employees:
268%
Number of
Vehicles:
215%
Demand
for
DATS:
234%
Forecast:

6. Shift Design Optimization
Part A:
Shift Design Optimization

7. Problem

8. Review the Effectiveness Overtime Reduction Perspective
Problem
Review the Effectiveness
of Part-Time Shifts
Maintain or Increase
Service Level with a
Greater Mix of Shifts
Overtime Reduction Perspective
in Reducing Daily
Operator Cost
Problem – Solution – Methodology – Analysis – Results - Risks
Problem – Solution – Methodology – Analysis – Results - Risks

9. Solution

10. Solution Problem – Solution – Methodology – Analysis – Results - Risks
Minimize Labour Cost
Increase Operator Utilization
Optimal Shift Design Model
Efficacy of Part-Time Shifts
Minimize labour cost and increase/maintain operator utilization
Lowering workers scheduled during peak times
Problem – Solution – Methodology – Analysis – Results - Risks
Problem – Solution – Methodology – Analysis – Results - Risks

11. Methodology

12. Methodology Shift Scheduling Model built in Excel
Queueing ToolPak (QTP)
Review Current Shift Schedule
Demand
Supply Constraints
Utilization
Input Various Shift Combinations
Cost Benefit Analysis
Excel based model
Queueing toolpak
Cost benefit analysis of current shift design vs potential shifts
Problem – Solution – Methodology – Analysis – Results - Risks
Problem – Solution – Methodology – Analysis – Results - Risks

13. Analysis

14. Daily Wheelchair Demand
Problem – Solution – Methodology – Analysis – Results - Risks

15. WC Demand vs. Scheduled Operators
1.9872…
Problem – Solution – Methodology – Analysis – Results - Risks

16. FT Shifts vs. PT Shifts
10.61% of staff PT.. 14 of 132
Problem – Solution – Methodology – Analysis – Results - Risks

17. Current Operator Utilization
Avg 56%, max 126, min 5%
Problem – Solution – Methodology – Analysis – Results - Risks

18. Monthly OT Costs ( )
Problem – Solution – Methodology – Analysis – Results - Risks

19. Results

20. Results Problem – Solution – Methodology – Analysis – Results - Risks
Current = 62 8, 17 10, 9PT ………FT <79… , 45, 8, 8 5…83tot
Problem – Solution – Methodology – Analysis – Results - Risks

21. Utilization: Old vs. New
Avg 56%, max 126, min 5%
Newbies- avg 61%. Not as steep of slope, smoothed
Problem – Solution – Methodology – Analysis – Results - Risks

22. Chosen Shift Blend
Maximum daily cost savings of $1093. savings of nearly $560 ($ yrly)…. Employ 10 less ppl on an avg basis…..Raise wages by 3.92%,
$14,258.28
$14,052.98
Problem – Solution – Methodology – Analysis – Results - Risks

23. Risks

24. Risks Higher utilization = More work for operators
decreased worker quality of life
Average trips made per shift will decrease
Employees prefer full-time shifts and overtime hours
Problem – Solution – Methodology – Analysis – Results - Risks

25. Part B:
Demand Forecasting

26. Problem

27. Problem Population of Edmonton: increasing changing Demand for DATS:
Resulting in:
planning challenges
fiscal constraints
Requiring:
forecast of future demand
forecast of impacts
Problem – Solution – Methodology – Analysis – Risks – Results

28. Solution

29. Solution Cost Containment Strategies Population of Edmonton:
age
gender
Disabled
Population
of Edmonton
Impacts on DATS:
vehicles
employees
operating
expenses
Demand
for DATS
Cost Containment
Strategies
Problem – Solution – Methodology – Analysis – Risks – Results

30. Methodology

31. Methodology Forecasting the Population of Edmonton
Disabled
Population of
Edmonton
Forecasting
the
Demand
for
DATS
Forecasting
the
Impacts on
DATS
Drivers of
Disability
Problem – Solution – Methodology – Analysis – Risks – Results

32. Analysis

33. Population of Edmonton
Growth:
total = 29%
65+ = 99%
Growth:
total = 1%
65+ = 4%
Growth:
total = 17%
65+ = 93%
Historical:
Forecast:
Problem – Solution – Methodology – Analysis – Risks – Results

34. Disabled Population of Edmonton
Growth:
total = 72%
65+ = 133%
Growth:
total = 62%
65+ = 125%
Growth:
total = 24%
65+ = 31%
Historical:
Forecast:
Problem – Solution – Methodology – Analysis – Risks - Results

35. Demand For DATS Growth: total = 256% 65+ = 460% Growth: total = 234%
65+ = 422%
Growth:
total = 167%
65+ = 302%
Historical:
Forecast:
Problem – Solution – Methodology – Analysis – Risks - Results

36. Number Of Vehicles Growth: 235% Growth: 29% Growth: 215%
Historical:
Forecast:
Problem – Solution – Methodology – Analysis – Risks - Results

37. Number Of Employees Growth: 293% Growth: 66% Growth: 268%
Historical:
Forecast:
Problem – Solution – Methodology – Analysis – Risks - Results

38. Total Operating Expenses
Growth:
215%
Growth:
52%
Growth:
197%
Historical:
Forecast:
Problem – Solution – Methodology – Analysis – Risks - Results

39. Risks

40. Demand For DATS Average Range = 29% Average Range = 29%
Problem – Solution – Methodology – Analysis – Risks – Results

41. Number Of Vehicles Average Range per Range in Demand = 0.95
Problem – Solution – Methodology – Analysis – Risks - Results

42. Number Of Employees Average Range per Range in Demand = 1.08
Problem – Solution – Methodology – Analysis – Risks - Results

43. Total Operating Expenses
Average Range
per
Range in Demand =
0.90
Average Range
per
Range in Demand =
0.90
Problem – Solution – Methodology – Analysis – Risks - Results

44. Results

45. Results Cost Containment Strategies Redirecting to Optimizing shift
schedule
Redirecting to
other assisted
transportation
services
Cost Containment
Strategies
Centralizing
pick-ups
and
drop-offs
Problem – Solution – Methodology – Analysis – Risks – Results

46. Key Findings

47. More PT Key Findings – Part A Less OT
Higher Utilization
With a greater balance of PT and FT shifts, OT will be reduced
Demand will be more effectively met from a cost perspective with the same service level

48. Key Findings – Part B Demand for DATS will grow significantly:
234% - 256%
Will cause significant impacts:
operating expenses increase by:
197% - 215%
Can be managed with cost containment strategies:
redirecting
optimizing
centralizing

49. Questions

50. Appendix

51. Appendix – Part A Threshold Time – 30 min Service Level - .99
Arrival Rate – Demand from 2 periods ahead averaged
Service Rate* – per hour per quarter hour
Queue Capacity – infinite
Lagg SIPP (stationary independent period by period) approach
MMS – Poisson Distribution, inter-event times (no exact info about next pickup)
*Service Rate = average(total passengers carried per operator / hours worked by operator)

52. Appendix – Part A ASSUMPTIONS Demand = SCHED_WC – CANCEL_WC
Demand broken down into weekday and weekend demand
No break time. Only current shift schedule includes split shifts
Tours 0.5hrs than actual length

53. Appendix – Part A
CONSTRAINTS
10 PT per day
Avg. Wage

54. Appendix – Part A
Labour Costs
Different Costs of Shift Combinations

55. Appendix – Part A

56. Appendix – Part A

57. Appendix – Part B: Data Limitations and Assumptions
Profiled only by age,
not also by gender
No data on disabled
population of
Edmonton
Data only on disabled
Alberta
for 2001 and
2006
Two data sets are
insufficient for
MVR
Two data sets only
allows forecasting
by SLR
DATS only tracks
registered clients,
not inquiries
Two data sets are
insufficient for MVR
Two data sets only
allows forecasting
by SLR
Assumed:
only cost driver is
number of clients
no intervention by
decision makers
no changes in
productivity