Calgary Commercial Movement Model

1
Calgary Commercial Movement Model

• Kevin Stefan, City of Calgary
• J.D. Hunt, University of Calgary
• Prepared for the 17th International EMME/2
  Conference
• Calgary, Alberta, Canada October 22-24, 2003

2
Calgary Commercial Movement Model

• Nature of Commercial Movement
• Basic properties
• Tour-based framework
• Model Structure
• Choice set tree
• Interface with RTM
• Java microsimulation
• Iterative processes

3
Why a Commercial Model?

• Represent all users of transportation systems
• Consumer surplus measures
• Emissions
• Volumes for design and microsimulation base
• Represent concerns and needs of business community

4
Commodity Flow Survey

• 24-hour business establishment travel diary
• Company information
• Industry, employees
• Fleet information
• Configuration, ownership, GVW
• Information at each stop made by each vehicle
• Location, time, purpose
• Type, amount and value of goods carried /
  services performed

5
Weekday Vehicle Trips
12
6
Weekday Vehicle Trips
7
Impact of Commercial Vehicles

• Higher value of time
• Travel concentrated in some areas
• Central Business District
• Industrial areas
• Commercial areas
• Increased influence in many factors
• Emissions
• Congestion and vehicle flow
• Pavement design

8
Commercial Fleet
Medium
Light
Heavy
Percentages of vehicle trips
9
Stop Purpose
10
Tour-based Framework

• Originated in personal models
• Considers tours rather than individual trips
• Uses additional information for decisions
• Previous stop locations
• Location of establishment (home base)
• Simulates each trip as tour progresses
• Similarities to commercial movement
• A number of clients scattered through city
• Efficient businesses will service them in tours

11
Establishment
12
Establishment
13
Establishment
14
Establishment
15
Establishment
16
Establishment
Lunch
17
Establishment
Lunch
18
Trip Purposes

• Goods
• Pick up goods
• Drop off goods
• Pick up and drop off goods
• Service
• Perform service
• Pick up supplies for performing service

19
Trip Purposes

• Other
• Vehicle repair or fueling
• Drivers personal stop (e.g. meal, snack, coffee,
  toilet, personal business)
• Stop for indirect business purpose (e.g. bank)
• Return
• Return to establishment
• Last stop of a tour

20
Trip Purposes
21
Fleet Allocators

• High trip volume by coordinated fleet
• A few specific generators
• Difficult to survey
• Less demand-sensitive (more typically a set route
  or coverage area)
• Outside the scope of this model a simpler (e.g.
  gravity) treatment

22
Fleet Allocators

• Mail and courier
• Garbage and recycling
• City parks and road maintenance
• Newspaper and flyer delivery

30
23
Model Structure
Tour Generation
Vehicle Type
Light
Medium
Heavy
Tour Type
Service
Goods
Other
Service
Goods
Other
Service
Goods
Other
Next Stop Type
Service
(Return)
(Return)
Service
(Return)
(Return)
Service
(Return)
(Return)
Other
Other
Other
Other
Other
Other
Goods
(Return)
Goods
(Return)
Goods
(Return)
Other
Other
Other
Next Stop Location
24
Model Structure
Tour Generation
Vehicle Type
Light
Medium
Heavy
Tour Type
Service
Goods
Other
Service
Goods
Other
Service
Goods
Other
Next Stop Type
Service
(Return)
(Return)
Service
(Return)
(Return)
Service
(Return)
(Return)
Other
Other
Other
Other
Other
Other
Goods
(Return)
Goods
(Return)
Goods
(Return)
Other
Other
Other
Next Stop Location
25
Model Structure
Logit choice model
Tour Generation
Vehicle Type
Light
Medium
Heavy
Tour Type
Service
Goods
Other
Service
Goods
Other
Service
Goods
Other
Next Stop Type
Service
(Return)
(Return)
Service
(Return)
(Return)
Service
(Return)
(Return)
Other
Other
Other
Other
Other
Other
Goods
(Return)
Goods
(Return)
Goods
(Return)
Other
Other
Other
Next Stop Location
26
Model Structure
Tour Generation
Vehicle Type
Light
Medium
Heavy
Tour Type
Service
Goods
Other
Service
Goods
Other
Service
Goods
Other
Next Stop Type
Service
(Return)
(Return)
Service
(Return)
(Return)
Service
(Return)
(Return)
Other
Other
Other
Other
Other
Other
Goods
(Return)
Goods
(Return)
Goods
(Return)
Other
Other
Other
Next Stop Location
27
Model Structure
Tour Generation
Vehicle Type
Light
Medium
Heavy
Tour Type
Service
Goods
Other
Service
Goods
Other
Service
Goods
Other
Next Stop Type
Service
(Return)
(Return)
Service
(Return)
(Return)
Service
(Return)
(Return)
Other
Other
Other
Other
Other
Other
Goods
(Return)
Goods
(Return)
Goods
(Return)
Other
Other
Other
Next Stop Location
28
Model Structure
Tour Generation
Vehicle Type
Light
Medium
Heavy
Tour Type
Service
Goods
Other
Service
Goods
Other
Service
Goods
Other
Next Stop Type
Service
(Return)
(Return)
Service
(Return)
(Return)
Service
(Return)
(Return)
Other
Other
Other
Other
Other
Other
Goods
(Return)
Goods
(Return)
Goods
(Return)
Other
Other
Other
Next Stop Location
29
Model Structure
Tour Generation
Vehicle Type
Light
Medium
Heavy
Tour Type
Service
Goods
Other
Service
Goods
Other
Service
Goods
Other
Next Stop Type
Service
(Return)
(Return)
Service
(Return)
(Return)
Service
(Return)
(Return)
Other
Other
Other
Other
Other
Other
Goods
(Return)
Goods
(Return)
Goods
(Return)
Other
Other
Other
Next Stop Location
30
Model Structure
Tour Generation
Vehicle Type
Light
Medium
Heavy
Tour Type
Service
Goods
Other
Service
Goods
Other
Service
Goods
Other
Next Stop Type
Service
(Return)
(Return)
Service
(Return)
(Return)
Service
(Return)
(Return)
Other
Other
Other
Other
Other
Other
Goods
(Return)
Goods
(Return)
Goods
(Return)
Other
Other
Other
Next Stop Location
31
Tour Generation

• Regression model
• Based on zonal information
• Population
• Employment by industry
• Accessibility factor
• Generates tours, not trips
• Policy-sensitive Responds to travel conditions
  (accessibility)

32
Model Structure
Tour Generation
Vehicle Type
Light
Medium
Heavy
Tour Type
Service
Goods
Other
Service
Goods
Other
Service
Goods
Other
Next Stop Type
Service
(Return)
(Return)
Service
(Return)
(Return)
Service
(Return)
(Return)
Other
Other
Other
Other
Other
Other
Goods
(Return)
Goods
(Return)
Goods
(Return)
Other
Other
Other
Next Stop Location
33
Vehicle Type

• Three classes of vehicles
• Based on establishment location
• Logsums from lower levels

Light
Medium
Heavy
34
Model Structure
Tour Generation
Vehicle Type
Light
Medium
Heavy
Tour Type
Service
Goods
Other
Service
Goods
Other
Service
Goods
Other
Next Stop Type
Service
(Return)
(Return)
Service
(Return)
(Return)
Service
(Return)
(Return)
Other
Other
Other
Other
Other
Other
Goods
(Return)
Goods
(Return)
Goods
(Return)
Other
Other
Other
Next Stop Location
35
Tour Type

• Based on establishment location and vehicle
  choice
• Observed very few tours mixing goods and service
  stops
• Other-only tours
• Vehicle repair
• Errand running

Vehicle Type
Goods (and other)
Other
Service (and other)
36
Model Structure
Tour Generation
Vehicle Type
Light
Medium
Heavy
Tour Type
Service
Goods
Other
Service
Goods
Other
Service
Goods
Other
Next Stop Type
Service
(Return)
(Return)
Service
(Return)
(Return)
Service
(Return)
(Return)
Other
Other
Other
Other
Other
Other
Goods
(Return)
Goods
(Return)
Goods
(Return)
Other
Other
Other
Next Stop Location
37
Next Stop Purpose
Tour Type Goods

• Different stop purposes affect stop location
• Different choice sets depending on tour type
• Other always option
• Return always option after first stop
• Return choice affects tour length

Goods
Return
Other
38
Next Stop Purpose
Tour Type Service

• Different stop purposes affect stop location
• Different choice sets depending on tour type
• Other always option
• Return always option after first stop
• Return choice affects tour length

Service
Return
Other
39
Next Stop Purpose
Tour Type Other

• Different stop purposes affect stop location
• Different choice sets depending on tour type
• Other always option
• Return always option after first stop
• Return choice affects tour length

Return
Other
40
Model Structure
Tour Generation
Vehicle Type
Light
Medium
Heavy
Tour Type
Service
Goods
Other
Service
Goods
Other
Service
Goods
Other
Next Stop Type
Service
(Return)
(Return)
Service
(Return)
(Return)
Service
(Return)
(Return)
Other
Other
Other
Other
Other
Other
Goods
(Return)
Goods
(Return)
Goods
(Return)
Other
Other
Other
Next Stop Location
41
Next Stop Location
Stop Type (Goods/Service/Other)

• Not needed if next stop purpose is return
• Choose from all 1447 zones in Calgary region
• Estimation uses 80 randomly selected zones

42
Model Structure
Tour Generation
Vehicle Type
Light
Medium
Heavy
Tour Type
Service
Goods
Other
Service
Goods
Other
Service
Goods
Other
Next Stop Type
Service
(Return)
(Return)
Service
(Return)
(Return)
Service
(Return)
(Return)
Other
Other
Other
Other
Other
Other
Goods
(Return)
Goods
(Return)
Goods
(Return)
Other
Other
Other
Next Stop Location
43
Model Features

• All levels feed into each other
• Whole model policy sensitive
• Truck routes for medium and heavy vehicles
• Calgary law Medium and heavy vehicles must
  travel on truck routes whenever possible
  (minimize travel off truck routes)

44
Truck Route Penalty
Dest.
Origin
45
Truck Route Penalty
Dest.
Origin
46
Truck Route Penalty
0.3
Dest.
2.0
0.3
0.2
1.5
1.4
1.2
0.9
3.1
1.2
Origin
3.0
Time 7.5 mins
47
Truck Route Penalty
100.3
Dest.
2.0
100.3
100.2
301.5
301.4
1.2
200.9
3.1
201.2
Origin
3.0
Time 408.1 minutes 408.1-(Int(408.1/100)100)8.1
minutes
48
Model Operation

• Trips generated one at a time in microsimulation
  module
• Microsimulation to get trip tables
• Java program external to EMME/2
• EMME/2 interface
• Reads data from main model database
• Generates tours in Java microsimulation (Monte
  Carlo)
• Trip table from microsimulation brought back into
  EMME/2 assigned to become volumes on links

49
Initial seed trip tables
A
equilibrium
ASSIGNMENT - get times costs
Get TRAVEL UTILITIES for each segment / mode /
time of day
Get daily TRIP DESTINATION ATTRACTIVENESS
Get DAILY COMPOSITE UTILITY
Get DAILY COMPOSITE UTILITY of ACCESIBILITY
Get DAILY PERSON TRIP GENERATION
Get DAILYPERSON TRIP DESTINATION choice
Commercial Movement Model
Calculate PERSON TRIP TABLES for each segment
Calculate mode shares for each segment / time of
day
22h - 2 iterations
Calculate vehicle trips by mode for all segments
A
B
Yes, do final iteration
No, calculate new trip tables
equilibrium
ASSIGNMENT - get times costs
Get TRAVEL UTILITIES for each segment / mode /
time of day
B
4h - 3 iterations
Calculate mode shares for each segment / time of
day
Calculate vehicle trips by mode for all segments
Final ASSIGNMENT all trips (auto, transit, walk,
bike)
Yes, re-calculate composite utilities / trip
tables
No, calculate new trip tables
50
Microsimulation Process
Tour Generation
Vehicle Type
Tour Type
Next Stop Purpose
Return
Service Goods Other
51
Microsimulation Process

• Tour starting in zone 340 (Central Industrial)

Tour Generation
Vehicle Type
Tour Type
Next Stop Purpose
Return
Service Goods Other
52
Microsimulation Process

• Tour starting in zone 340 (Central Industrial)
• Light vehicle

Tour Generation
Vehicle Type
Tour Type
Next Stop Purpose
Return
Service Goods Other
53
Microsimulation Process

• Tour starting in zone 340 (Central Industrial)
• Light vehicle
• Service (and other) tour

Tour Generation
Vehicle Type
Tour Type
Next Stop Purpose
Return
Service Goods Other
54
Microsimulation Process

• Tour starting in zone 340 (Central Industrial)
• Light vehicle
• Service (and other) tour
• Service stop

Tour Generation
Vehicle Type
Tour Type
Next Stop Purpose
Return
Service Goods Other
55
Microsimulation Process

• Tour starting in zone 340 (Central Industrial)
• Light vehicle
• Service (and other) tour
• Service, 211 (Stampede)

Tour Generation
Vehicle Type
Tour Type
Next Stop Purpose
Return
Service Goods Other
56
Microsimulation Process

• Tour starting in zone 340 (Central Industrial)
• Light vehicle
• Service (and other) tour
• Service, 211 (Stampede)
• Service stop

Tour Generation
Vehicle Type
Tour Type
Next Stop Purpose
Return
Service Goods Other
57
Microsimulation Process

• Tour starting in zone 340 (Central Industrial)
• Light vehicle
• Service (and other) tour
• Service, 211 (Stampede)
• Service, 209 (Apartment)

Tour Generation
Vehicle Type
Tour Type
Next Stop Purpose
Return
Service Goods Other
58
Microsimulation Process

• Tour starting in zone 340 (Central Industrial)
• Light vehicle
• Service (and other) tour
• Service, 211 (Stampede)
• Service, 209 (Apartment)
• Other stop

Tour Generation
Vehicle Type
Tour Type
Next Stop Purpose
Return
Service Goods Other
59
Microsimulation Process

• Tour starting in zone 340 (Central Industrial)
• Light vehicle
• Service (and other) tour
• Service, 211 (Stampede)
• Service, 209 (Apartment)
• Other, 2205 (Marathon rest.)

Tour Generation
Vehicle Type
Tour Type
Next Stop Purpose
Return
Service Goods Other
60
Microsimulation Process

• Tour starting in zone 340 (Central Industrial)
• Light vehicle
• Service (and other) tour
• Service, 211 (Stampede)
• Service, 209 (Apartment)
• Other, 2205 (Marathon rest.)
• Service stop

Tour Generation
Vehicle Type
Tour Type
Next Stop Purpose
Return
Service Goods Other
61
Microsimulation Process

• Tour starting in zone 340 (Central Industrial)
• Light vehicle
• Service (and other) tour
• Service, 211 (Stampede)
• Service, 209 (Apartment)
• Other, 2205 (Marathon rest.)
• Service, 2312 (North Hill Mall)

Tour Generation
Vehicle Type
Tour Type
Next Stop Purpose
Return
Service Goods Other
62
Microsimulation Process

• Tour starting in zone 340 (Central Industrial)
• Light vehicle
• Service (and other) tour
• Service, 211 (Stampede)
• Service, 209 (Apartment)
• Other, 2205 (Marathon rest.)
• Service, 2312 (North Hill Mall)
• Return to establishment

Tour Generation
Vehicle Type
Tour Type
Next Stop Purpose
Return
Service Goods Other
63
Microsimulation Process

• Tour starting in zone 340 (Central Industrial)
• Light vehicle
• Service (and other) tour
• Service, 211 (Stampede)
• Service, 209 (Apartment)
• Other, 2205 (Marathon rest.)
• Service, 2312 (North Hill Mall)
• Return to establishment, 340

Tour Generation
Vehicle Type
Tour Type
Next Stop Purpose
Return
Service Goods Other
64
Microsimulation Process

• Tour starting in zone 340 (Central Industrial)
• Light vehicle
• Service (and other) tour
• Service, 211 (Stampede)
• Service, 209 (Apartment)
• Other, 2205 (Marathon rest.)
• Service, 2312 (North Hill Mall)
• Return to establishment, 340
• Tour starting in zone 2604 (NW residential)

Tour Generation
Vehicle Type
Tour Type
Next Stop Purpose
Return
Service Goods Other
65
Microsimulation and Convergence

• Potential problem
• EMME/2 is deterministic
• Microsimulation is stochastic
• Random kick to EMME/2 traffic volumes each
  iteration
• Travel times change for each output
• Commercial vehicle flows generated then assigned
  no feedback or convergence

66
Travel time updating
Personal volumes

• Update travel times after a microsim. run
• Use as input to another microsim. run
• Iterative

EMME/2 Databank
Travel time O-D skims
Commercial trip table
Microsimulation
67
Travel time updating
Personal volumes

• Update travel times after a microsim. run
• Use as input to another microsim. run
• Iterative

EMME/2 Databank
Updated Travel times
Commercial trip table
Microsimulation
68
Deterministic vs. Stochastic
200 vehicles
126
125
69
Deterministic vs. Stochastic
Model Run 1
EMME/2 112.3
Microsimulation 125
126
125
EMME/2 87.7
Microsimulation 75
70
Deterministic vs. Stochastic
Model Run 1 2
EMME/2 112.3 112.3
Microsimulation 125 99
126
125
EMME/2 87.7 87.7
Microsimulation 75 101
71
Deterministic vs. Stochastic
Model Run 1 2 3
EMME/2 112.3 112.3 112.3
Microsimulation 125 99 117
126
125
EMME/2 87.7 87.7 87.7
Microsimulation 75 101 83
72
Deterministic vs. Stochastic
Model Run 1 2 3
EMME/2 112.3 112.3 112.3
Microsimulation 125 99 117
126
125
EMME/2 87.7 87.7 87.7
Microsimulation 75 101 83
73
Combining Deterministic and Stochastic Techniques

• Multiple runs of microsimulation model with same
  inputs will yield a range of output trip tables
• Run microsimulation multiple times and average
  the trip table results
• Fluctuations of microsimulation will tend to
  cancel each other out
• Outputs will be more predictable

74
Commercial Vehicle Model
Info from EMME/2 Databank
75
Status

• Data cleaning and analysis done
• Preliminary estimations May-Aug 2003
• Model design done
• Choice data sets under construction
• Truck route network coded and assignment working
• Microsimulator done

76
Summary

• Commercial vehicle movement is significant and
  complex
• Commercial transportation is important
• Interplay of vehicle types, truck routes
• Variety of stop purposes
• Tour-based methodology is appropriate
• Reflects efficient servicing of various clients
• Includes information about previous stops

77
Summary

• Nested logit choice model
• Four choice levels plus generation regression
• Provides for information and policy effect at all
  levels
• Requires truck route multiclass assignment
• Still under development
• Java-based microsimulation
• Interfaces with EMME/2 to obtain data and supply
  trip tables
• Iterative approach to solve convergence issues

78
(No Transcript)