CIE4801 Transportation and spatial modelling (Uncongested ...
Transcript of CIE4801 Transportation and spatial modelling (Uncongested ...
![Page 1: CIE4801 Transportation and spatial modelling (Uncongested ...](https://reader034.fdocuments.in/reader034/viewer/2022051600/627fa84b1a7c80231a2ed1a2/html5/thumbnails/1.jpg)
31-08-18
Challenge the future
DelftUniversity ofTechnology
CIE4801 Transportation and spatial modelling
Rob van Nes, Transport & Planning
(Uncongested) Assignment
![Page 2: CIE4801 Transportation and spatial modelling (Uncongested ...](https://reader034.fdocuments.in/reader034/viewer/2022051600/627fa84b1a7c80231a2ed1a2/html5/thumbnails/2.jpg)
2CIE4801: (Uncongested) Assignment
2.Networks
![Page 3: CIE4801 Transportation and spatial modelling (Uncongested ...](https://reader034.fdocuments.in/reader034/viewer/2022051600/627fa84b1a7c80231a2ed1a2/html5/thumbnails/3.jpg)
3CIE4801: (Uncongested) Assignment
Constructing a transport network
Given a map of the study area, how to represent the infrastructure and thetravel demand in a model?
![Page 4: CIE4801 Transportation and spatial modelling (Uncongested ...](https://reader034.fdocuments.in/reader034/viewer/2022051600/627fa84b1a7c80231a2ed1a2/html5/thumbnails/4.jpg)
4CIE4801: (Uncongested) Assignment
Network attributes
node• x-coordinate• y-coordinate
centroid node• zonal data• origin/destination
link• node-from• node-to• length• maximum speed• number of lanes• capacity
![Page 5: CIE4801 Transportation and spatial modelling (Uncongested ...](https://reader034.fdocuments.in/reader034/viewer/2022051600/627fa84b1a7c80231a2ed1a2/html5/thumbnails/5.jpg)
5CIE4801: (Uncongested) Assignment
Links and junctions
=
Junction with allturns allowed
Junction with no left turns allowedUsing techniques like this you can
model various types of intersections
![Page 6: CIE4801 Transportation and spatial modelling (Uncongested ...](https://reader034.fdocuments.in/reader034/viewer/2022051600/627fa84b1a7c80231a2ed1a2/html5/thumbnails/6.jpg)
6CIE4801: (Uncongested) Assignment
Define zones and select roads
![Page 7: CIE4801 Transportation and spatial modelling (Uncongested ...](https://reader034.fdocuments.in/reader034/viewer/2022051600/627fa84b1a7c80231a2ed1a2/html5/thumbnails/7.jpg)
7CIE4801: (Uncongested) Assignment
Select links
4321 nodes1217 zones
826 nodes 170 zones
204 nodes 36 zones
centroidnodereal linkconnector
How many zones / nodes / links?
• depends on the application
• include one network level lower thanthe network level you’re interested in
modelling=
the art of leaving things out
![Page 8: CIE4801 Transportation and spatial modelling (Uncongested ...](https://reader034.fdocuments.in/reader034/viewer/2022051600/627fa84b1a7c80231a2ed1a2/html5/thumbnails/8.jpg)
8CIE4801: (Uncongested) Assignment
Which roads should be included?
![Page 9: CIE4801 Transportation and spatial modelling (Uncongested ...](https://reader034.fdocuments.in/reader034/viewer/2022051600/627fa84b1a7c80231a2ed1a2/html5/thumbnails/9.jpg)
9CIE4801: (Uncongested) Assignment
Urban or regional model?
Regional Urban
![Page 10: CIE4801 Transportation and spatial modelling (Uncongested ...](https://reader034.fdocuments.in/reader034/viewer/2022051600/627fa84b1a7c80231a2ed1a2/html5/thumbnails/10.jpg)
10CIE4801: (Uncongested) Assignment
Example car network regional model
![Page 11: CIE4801 Transportation and spatial modelling (Uncongested ...](https://reader034.fdocuments.in/reader034/viewer/2022051600/627fa84b1a7c80231a2ed1a2/html5/thumbnails/11.jpg)
11CIE4801: (Uncongested) Assignment
Important issue
• Connecting the zones to the network:
• Single connector or multiple connectors?
• Connecting to which type of node/link?
• Choices have major consequences for the assignment to the network!
• Preferably multiple connectors and connected to lower level network in your model
![Page 12: CIE4801 Transportation and spatial modelling (Uncongested ...](https://reader034.fdocuments.in/reader034/viewer/2022051600/627fa84b1a7c80231a2ed1a2/html5/thumbnails/12.jpg)
12CIE4801: (Uncongested) Assignment
3.1Main building block: shortest path
![Page 13: CIE4801 Transportation and spatial modelling (Uncongested ...](https://reader034.fdocuments.in/reader034/viewer/2022051600/627fa84b1a7c80231a2ed1a2/html5/thumbnails/13.jpg)
13CIE4801: (Uncongested) Assignment
Shortest Path algorithms
• “Oldies”• Moore (1959)• Dijkstra (1959)• (Floyd-Warschall (1962))
• Still a topic for research
![Page 14: CIE4801 Transportation and spatial modelling (Uncongested ...](https://reader034.fdocuments.in/reader034/viewer/2022051600/627fa84b1a7c80231a2ed1a2/html5/thumbnails/14.jpg)
14CIE4801: (Uncongested) Assignment
Main concept for ‘tree algorithms’(Moore and Dijkstra)
• For all nodes• Set travel time tt(x) to and set the back node bn(x) to 0
• For the origin i set time to 0 and back node to -1• Node i is the first active node a• Select all links (a,j) and check travel times
• If tt(a)+time(a,j) < tt(j)tt(j)=tt(a)+time(a,j) and bn(j)=a and node j becomes an active node
• Node a is no longer active• Select a new active node from the stack of active nodes and
repeat previous step until there are no active nodes left
• For Dijkstra: select the link having the lowest travel time and select the active node having the lowest travel time
¥
![Page 15: CIE4801 Transportation and spatial modelling (Uncongested ...](https://reader034.fdocuments.in/reader034/viewer/2022051600/627fa84b1a7c80231a2ed1a2/html5/thumbnails/15.jpg)
15CIE4801: (Uncongested) Assignment
2 1
3
2
1 1
2
1
1
5
6
4
Shortest paths
![Page 16: CIE4801 Transportation and spatial modelling (Uncongested ...](https://reader034.fdocuments.in/reader034/viewer/2022051600/627fa84b1a7c80231a2ed1a2/html5/thumbnails/16.jpg)
16CIE4801: (Uncongested) Assignment
2 1
3
2
1 1
2
1
1
5
6
4
20
Shortest paths
![Page 17: CIE4801 Transportation and spatial modelling (Uncongested ...](https://reader034.fdocuments.in/reader034/viewer/2022051600/627fa84b1a7c80231a2ed1a2/html5/thumbnails/17.jpg)
17CIE4801: (Uncongested) Assignment
2 1
3
2
1 1
2
1
1
5
6
4
30 2
Shortest paths
![Page 18: CIE4801 Transportation and spatial modelling (Uncongested ...](https://reader034.fdocuments.in/reader034/viewer/2022051600/627fa84b1a7c80231a2ed1a2/html5/thumbnails/18.jpg)
18CIE4801: (Uncongested) Assignment
2 1
3
2
1 1
2
1
1
5
6
4
3
3
2
Shortest paths
![Page 19: CIE4801 Transportation and spatial modelling (Uncongested ...](https://reader034.fdocuments.in/reader034/viewer/2022051600/627fa84b1a7c80231a2ed1a2/html5/thumbnails/19.jpg)
19CIE4801: (Uncongested) Assignment
2 1
3
2
1 1
2
1
1
5
6
4 43
2
Shortest paths
![Page 20: CIE4801 Transportation and spatial modelling (Uncongested ...](https://reader034.fdocuments.in/reader034/viewer/2022051600/627fa84b1a7c80231a2ed1a2/html5/thumbnails/20.jpg)
20CIE4801: (Uncongested) Assignment
2 1
3
2
1 1
2
1
1
5
6
4 4
5
3
2
Shortest paths
![Page 21: CIE4801 Transportation and spatial modelling (Uncongested ...](https://reader034.fdocuments.in/reader034/viewer/2022051600/627fa84b1a7c80231a2ed1a2/html5/thumbnails/21.jpg)
21CIE4801: (Uncongested) Assignment
2 1
3
2
1 1
2
1
1
5
6
4
65
43
2
Shortest paths
![Page 22: CIE4801 Transportation and spatial modelling (Uncongested ...](https://reader034.fdocuments.in/reader034/viewer/2022051600/627fa84b1a7c80231a2ed1a2/html5/thumbnails/22.jpg)
22CIE4801: (Uncongested) Assignment
2 1
3
2
1 1
2
1
1
5
6
4
66
43
2
Shortest paths
![Page 23: CIE4801 Transportation and spatial modelling (Uncongested ...](https://reader034.fdocuments.in/reader034/viewer/2022051600/627fa84b1a7c80231a2ed1a2/html5/thumbnails/23.jpg)
23CIE4801: (Uncongested) Assignment
2 1
3
2
1
2
1
1
Shortest path treeContains shortest paths to all nodes from a certain origin
Shortest paths
![Page 24: CIE4801 Transportation and spatial modelling (Uncongested ...](https://reader034.fdocuments.in/reader034/viewer/2022051600/627fa84b1a7c80231a2ed1a2/html5/thumbnails/24.jpg)
24CIE4801: (Uncongested) Assignment
Main concept for matrix algorithm(Floyd-Warschall)
• Create two matrices from all nodes to all nodes• tt(i,j) = t(i,j) for all links (i,j) or tt(i,j) = • bn(i,j)=i for all links (i,j)
• For every node k check travel times• If tt(i,k) + tt(k,j) < tt(i,j)
tt(i,j) = tt(i,k) + tt(k,j) and bn(i,j)=bn(k,j)
• Repeat previous step until no changes are made
• Interesting option for public transport
¥
![Page 25: CIE4801 Transportation and spatial modelling (Uncongested ...](https://reader034.fdocuments.in/reader034/viewer/2022051600/627fa84b1a7c80231a2ed1a2/html5/thumbnails/25.jpg)
25CIE4801: (Uncongested) Assignment
3.2Shortest path representation
![Page 26: CIE4801 Transportation and spatial modelling (Uncongested ...](https://reader034.fdocuments.in/reader034/viewer/2022051600/627fa84b1a7c80231a2ed1a2/html5/thumbnails/26.jpg)
26CIE4801: (Uncongested) Assignment
11
4
7
2
5
8
3
6
9
2
3 4 56 7
8 9 1011 12
Shortest route from node 7 to node 5:
node 7 à node 4 à node 1 à node 2 à node 5orlink 8 à link 3 à link 1 à link 4
Representation of shortest paths
![Page 27: CIE4801 Transportation and spatial modelling (Uncongested ...](https://reader034.fdocuments.in/reader034/viewer/2022051600/627fa84b1a7c80231a2ed1a2/html5/thumbnails/27.jpg)
27CIE4801: (Uncongested) Assignment
11
4
7
2
5
8
3
6
9
2
3 4 56 7
8 9 1011 12
Very compact, less practical:Back node representation
4 1 6 7 2 9 -1 7 8
node
1no
de 2
node
3no
de 4
node
5no
de 6
node
7no
de 8
node
9
route “7 à x”
Number for elements for storing allshortest routes in the network:
9 9 81N N´ = × =
Representation of shortest paths
![Page 28: CIE4801 Transportation and spatial modelling (Uncongested ...](https://reader034.fdocuments.in/reader034/viewer/2022051600/627fa84b1a7c80231a2ed1a2/html5/thumbnails/28.jpg)
28CIE4801: (Uncongested) Assignment
11
4
7
2
5
8
3
6
9
2
3 4 56 7
8 9 1011 12
Very compact, more practical:Back node representationPlus link numbers 4 1 6 7 2 9 -1 7 8
node
1no
de 2
node
3no
de 4
node
5no
de 6
node
7no
de 8
node
9
route “7 à x”
Number for elements for storing allshortest routes in the network:2 2 9 9 162N N´ ´ = × × =
Representation of shortest paths
3 1 5 8 4 10 -1 11 12
![Page 29: CIE4801 Transportation and spatial modelling (Uncongested ...](https://reader034.fdocuments.in/reader034/viewer/2022051600/627fa84b1a7c80231a2ed1a2/html5/thumbnails/29.jpg)
29CIE4801: (Uncongested) Assignment
Assignment algorithm using back nodes
• Set all link flows equal to 0• For each origin
• Determine shortest path tree• For each destination
• Backtrack the route and add the OD-flow to the link flow of each link in the route
• Advantage: fast, efficient memory usage
![Page 30: CIE4801 Transportation and spatial modelling (Uncongested ...](https://reader034.fdocuments.in/reader034/viewer/2022051600/627fa84b1a7c80231a2ed1a2/html5/thumbnails/30.jpg)
30CIE4801: (Uncongested) Assignment
4.1Introduction to assignment
![Page 31: CIE4801 Transportation and spatial modelling (Uncongested ...](https://reader034.fdocuments.in/reader034/viewer/2022051600/627fa84b1a7c80231a2ed1a2/html5/thumbnails/31.jpg)
31CIE4801: (Uncongested) Assignment
What do we want to know?
• Often the main goal of a modelling study
• Flows on links or lines• Bottlenecks• Travel times
• This is where demand really meets supply
![Page 32: CIE4801 Transportation and spatial modelling (Uncongested ...](https://reader034.fdocuments.in/reader034/viewer/2022051600/627fa84b1a7c80231a2ed1a2/html5/thumbnails/32.jpg)
32CIE4801: (Uncongested) Assignment
Introduction to assignment
assignmentmodel
travel demand
infrastructure supply
flows
![Page 33: CIE4801 Transportation and spatial modelling (Uncongested ...](https://reader034.fdocuments.in/reader034/viewer/2022051600/627fa84b1a7c80231a2ed1a2/html5/thumbnails/33.jpg)
33CIE4801: (Uncongested) Assignment
Solving the assignment problem
1O
2O
1D
2D
2 1
1
3
1
2
2
11
1
2
11O
2O
1D 2D
10 5
8 3
OD matrix network
Given the travel demand and the infrastructure supply,determine assignment using the shortest paths.
![Page 34: CIE4801 Transportation and spatial modelling (Uncongested ...](https://reader034.fdocuments.in/reader034/viewer/2022051600/627fa84b1a7c80231a2ed1a2/html5/thumbnails/34.jpg)
34CIE4801: (Uncongested) Assignment
Solving the assignment problem
1O
2O
1D
2D
1O
2O
1D 2D
10 5
8 3
OD matrix network
10 10
5 5
55
For the first origin, find all shortest pathsto all destinations, and assign the correspondingtrips to the shortest paths.
![Page 35: CIE4801 Transportation and spatial modelling (Uncongested ...](https://reader034.fdocuments.in/reader034/viewer/2022051600/627fa84b1a7c80231a2ed1a2/html5/thumbnails/35.jpg)
35CIE4801: (Uncongested) Assignment
Solving the assignment problem
1O
2O
1D
2D
1O
2O
1D 2D
10 5
8 3
OD matrix network
11 38
88
For the second origin, find all shortest pathsto all destinations, and assign the correspondingtrips to the shortest paths.
![Page 36: CIE4801 Transportation and spatial modelling (Uncongested ...](https://reader034.fdocuments.in/reader034/viewer/2022051600/627fa84b1a7c80231a2ed1a2/html5/thumbnails/36.jpg)
36CIE4801: (Uncongested) Assignment
Solving the assignment problem
1O
2O
1D
2D
1O
2O
1D 2D
10 5
8 3
OD matrix network11 8
58
810 105
50 0
Combine all assignments between all OD pairs,and compute the total link flows.
0
800 0
0
00
0 0
00
![Page 37: CIE4801 Transportation and spatial modelling (Uncongested ...](https://reader034.fdocuments.in/reader034/viewer/2022051600/627fa84b1a7c80231a2ed1a2/html5/thumbnails/37.jpg)
37CIE4801: (Uncongested) Assignment
Assignment algorithm using back nodes
• Set all link flows equal to 0• For each origin
• Determine shortest path tree• For each destination
• Backtrack the route and add the OD-flow to the link flow of each link in the route
• Advantage: fast, efficient memory usage
![Page 38: CIE4801 Transportation and spatial modelling (Uncongested ...](https://reader034.fdocuments.in/reader034/viewer/2022051600/627fa84b1a7c80231a2ed1a2/html5/thumbnails/38.jpg)
38CIE4801: (Uncongested) Assignment
This was how you might do it
• Does it capture the relevant phenomena for the assignment?
• Differences in perception between users• Single route per OD or multiple routes• Multiple route concept implies route choice modelling
• Effect of congestion• Included or excluded
![Page 39: CIE4801 Transportation and spatial modelling (Uncongested ...](https://reader034.fdocuments.in/reader034/viewer/2022051600/627fa84b1a7c80231a2ed1a2/html5/thumbnails/39.jpg)
39CIE4801: (Uncongested) Assignment
Main assignment types
Congestion effect modelled?
No Yes
All-or-nothingassignment
Deterministicuser-equilibrium
assignment
Stochasticassignment
Stochasticuser-equilibrium
assignment
Multiple routesmodelled?
No
Yes
![Page 40: CIE4801 Transportation and spatial modelling (Uncongested ...](https://reader034.fdocuments.in/reader034/viewer/2022051600/627fa84b1a7c80231a2ed1a2/html5/thumbnails/40.jpg)
40CIE4801: (Uncongested) Assignment
4.2All-or-Nothing assignment
![Page 41: CIE4801 Transportation and spatial modelling (Uncongested ...](https://reader034.fdocuments.in/reader034/viewer/2022051600/627fa84b1a7c80231a2ed1a2/html5/thumbnails/41.jpg)
41CIE4801: (Uncongested) Assignment
All-or-Nothing (AON) assignment
Assumptions:• All traffic between an OD pair is assigned to just one route
(usually the ‘shortest’ route (in terms of cost (min, €, km)))• There is no congestion
Topics• Mathematical formulation• Some comments
![Page 42: CIE4801 Transportation and spatial modelling (Uncongested ...](https://reader034.fdocuments.in/reader034/viewer/2022051600/627fa84b1a7c80231a2ed1a2/html5/thumbnails/42.jpg)
42CIE4801: (Uncongested) Assignment
AON: Formulated as an optimisationproblem (route based)
ijr
ijr
ij
T
t
T
=
=
=
number of OD trips from i to j taking route r
travel time on route r from i to j
number of OD trips from i to j
minimize the total travel time experienced by all travellers
ijr ijrT T=å flow conservation
0ijrT ³ non-negativities
minijr
ijr ijrT i j rZ t T=ååå
s.t. Note that using this formulation you could have multiple routes having a similar travel time!
![Page 43: CIE4801 Transportation and spatial modelling (Uncongested ...](https://reader034.fdocuments.in/reader034/viewer/2022051600/627fa84b1a7c80231a2ed1a2/html5/thumbnails/43.jpg)
43CIE4801: (Uncongested) Assignment
asa
a
ms
q
qtT
=
=
=
=
traffic flow on link a
traffic flow on link a with destination s
travel time on link a
number of OD trips from node m to node s (m is an origin)
minimize the total travel time experienced by all travellersmin
aa aq a
Z t q=ås.t.
;sa a
sq q=å s s
ms a aa M a M
T q q m s- +Î Î
+ = " ¹å å flow conservation
0saq ³ non-negativities
AON: Formulated as an optimisationproblem (link based)
M-: incoming; M+: outgoing
![Page 44: CIE4801 Transportation and spatial modelling (Uncongested ...](https://reader034.fdocuments.in/reader034/viewer/2022051600/627fa84b1a7c80231a2ed1a2/html5/thumbnails/44.jpg)
44CIE4801: (Uncongested) Assignment
AON-assignment: some comments
• Very fast method• Easy to interpret
• Misses essential concepts• Very sensitive for small changes
• Acceptable if you have many zones?
• Note that AON of software package A might differ from AON of software package B
![Page 45: CIE4801 Transportation and spatial modelling (Uncongested ...](https://reader034.fdocuments.in/reader034/viewer/2022051600/627fa84b1a7c80231a2ed1a2/html5/thumbnails/45.jpg)
45CIE4801: (Uncongested) Assignment
4.3Multiple routes: route choice(stochastic assignment)
![Page 46: CIE4801 Transportation and spatial modelling (Uncongested ...](https://reader034.fdocuments.in/reader034/viewer/2022051600/627fa84b1a7c80231a2ed1a2/html5/thumbnails/46.jpg)
46CIE4801: (Uncongested) Assignment
Stochastic assignment
Shortest path: All-or-nothing assignment:For each OD pair, all travellers take the shortest route
Multiple routes: Stochastic assignmentFor each OD pair, all travellers take their perceivedshortest route
1t 2t 3t
1t2t
3t
travel time
prob
abilit
y
![Page 47: CIE4801 Transportation and spatial modelling (Uncongested ...](https://reader034.fdocuments.in/reader034/viewer/2022051600/627fa84b1a7c80231a2ed1a2/html5/thumbnails/47.jpg)
47CIE4801: (Uncongested) Assignment
Discrete choice theory( for all )i i jp P U U j= ³
( for all )i i j jP V V je e= + ³ +
Probit-model
.ip =!ee
i
j
V
i V
j
pb
b=å
Logit-model
If ’s are all extreme value distributed (independent, with scale parameter ),
ie
β
If ’s are all normally distributed (independent),
ie
Easy to solve Can only be solvedby simulation
( for all )i j j iP V V je e= - ³ -
![Page 48: CIE4801 Transportation and spatial modelling (Uncongested ...](https://reader034.fdocuments.in/reader034/viewer/2022051600/627fa84b1a7c80231a2ed1a2/html5/thumbnails/48.jpg)
48CIE4801: (Uncongested) Assignment
Probit and logit in transport models
• Note that previous slide is independent on type of choice
• Modelling practice:• Demand modelling: dominated by logit• Assignment: Probit or logit
• In case of assignment the link is the basic element, thus ideally the error term should be defined at link level
• In this case normal distribution is the most likely assumption• Thus Probit
![Page 49: CIE4801 Transportation and spatial modelling (Uncongested ...](https://reader034.fdocuments.in/reader034/viewer/2022051600/627fa84b1a7c80231a2ed1a2/html5/thumbnails/49.jpg)
49CIE4801: (Uncongested) Assignment
How is the probit assignment solved?
• Iteratively
• In each iteration a specific network state is consideredi.e. a network having specific link times
• These link times are sampled from the distributions that are assumed at link level
• Note that overlapping routes in a specific network state are consistent
![Page 50: CIE4801 Transportation and spatial modelling (Uncongested ...](https://reader034.fdocuments.in/reader034/viewer/2022051600/627fa84b1a7c80231a2ed1a2/html5/thumbnails/50.jpg)
50CIE4801: (Uncongested) Assignment
Example
t1 = N (10,2) t2 = N (5,0.1)
t3 = N (5,0.1)
State 1
State 2
t1 = 10.85 t2 = 4.98
t3 = 5.19
t1 = 8.57 t2 = 4.89
t3 = 4.99
Route 1=15.83Route 2=16.04
Route 1=13.46Route 2=13.56
![Page 51: CIE4801 Transportation and spatial modelling (Uncongested ...](https://reader034.fdocuments.in/reader034/viewer/2022051600/627fa84b1a7c80231a2ed1a2/html5/thumbnails/51.jpg)
51CIE4801: (Uncongested) Assignment
Scaling of the error term
( )a a at L= + × Qt z Dispersion parameterSize of link
Random variable e.g. (0,1)aL a
N
Q ==
=z
The error term is related to the ‘size' of link
Size could be the length or free flow travel time of link
![Page 52: CIE4801 Transportation and spatial modelling (Uncongested ...](https://reader034.fdocuments.in/reader034/viewer/2022051600/627fa84b1a7c80231a2ed1a2/html5/thumbnails/52.jpg)
52CIE4801: (Uncongested) Assignment
Probit assignment at link level
No explicit analytical expression, but can be computed by simulation.
10
55
5
62%31%
31%
38%
For n=1 to N do1. 2. Compute shortest route3. If shortest route is alternative r
then
0 0: ( ) , (0,1)a a at t t N= + Q z z !
: 1/ijr ijr Nf f= +
![Page 53: CIE4801 Transportation and spatial modelling (Uncongested ...](https://reader034.fdocuments.in/reader034/viewer/2022051600/627fa84b1a7c80231a2ed1a2/html5/thumbnails/53.jpg)
53CIE4801: (Uncongested) Assignment
Stochastic assignment
9.3600 9.5860 10.476610.9212 7.9679 9.13409.8664 8.5351 9.06839.5078 11.0549 11.23269.1196 10.3319 10.01109.8607 10.5413 9.22579.6406 10.8822 11.0726
10.4757 9.9555 8.007311.4155 9.5061 9.193710.3215 9.8450 10.7971
10 iterations:route 1 route 2 route 3
40% 30% 30%
Probit assignment:
10
55
5
![Page 54: CIE4801 Transportation and spatial modelling (Uncongested ...](https://reader034.fdocuments.in/reader034/viewer/2022051600/627fa84b1a7c80231a2ed1a2/html5/thumbnails/54.jpg)
54CIE4801: (Uncongested) Assignment
Stochastic assignment at route level
Assume that z is Extreme value distributed:
exp( )exp( )
ijrijr
ijpp
tt
bf
b-
=-å
(Logit)
10
55
5
But ... logit assignment cannot deal properly withoverlapping routes!
All 3 route alternatives: 10 min.
33%
67%33%
33%
![Page 55: CIE4801 Transportation and spatial modelling (Uncongested ...](https://reader034.fdocuments.in/reader034/viewer/2022051600/627fa84b1a7c80231a2ed1a2/html5/thumbnails/55.jpg)
55CIE4801: (Uncongested) Assignment
Stochastic assignment at route level
If there are simple overlapping routes, we can use thenested logit model
10
55
5
39%
61%30%
30%
10
5 4.5
22
2
1 ln (exp( ))
( 1.5)
aa
t tbb
b
= - -
=
å
![Page 56: CIE4801 Transportation and spatial modelling (Uncongested ...](https://reader034.fdocuments.in/reader034/viewer/2022051600/627fa84b1a7c80231a2ed1a2/html5/thumbnails/56.jpg)
56CIE4801: (Uncongested) Assignment
Stochastic assignment: comments
• In which cases would it be a realistic assignment?• Networks without congestion
• Bicycle, 24-hour period (?)
• How many iterations?• Old models: maximum of e.g. 6 (Burrell)• Theoretically: iterate until convergence
• Normal distribution?• Only positive travel times => e.g. Gamma distribution
![Page 57: CIE4801 Transportation and spatial modelling (Uncongested ...](https://reader034.fdocuments.in/reader034/viewer/2022051600/627fa84b1a7c80231a2ed1a2/html5/thumbnails/57.jpg)
57CIE4801: (Uncongested) Assignment
5.Assignment map and Path based approach
![Page 58: CIE4801 Transportation and spatial modelling (Uncongested ...](https://reader034.fdocuments.in/reader034/viewer/2022051600/627fa84b1a7c80231a2ed1a2/html5/thumbnails/58.jpg)
58CIE4801: (Uncongested) Assignment
11
4
7
2
5
8
3
6
9
2
3 4 56 7
8 9 1011 12
link 1 0 1 0 0 1 0 0 0link 2 0 0 0 0 0 0 0 0link 3 1 1 0 0 1 0 0 0link 4 0 0 0 0 1 0 0 0link 5 0 0 1 0 0 0 0 0link 6 0 0 0 0 0 0 0 0link 7 0 0 0 0 0 0 0 0link 8 1 1 0 1 1 0 0 0link 9 0 0 0 0 0 0 0 0link 10 0 0 1 0 0 1 0 0link 11 0 0 1 0 0 1 1 1link 12 0 0 1 0 0 1 0 1
rout
e 7à
1ro
ute
7à2
rout
e 7à
3ro
ute
7à4
rout
e 7à
5ro
ute
7à6
rout
e 7à
8ro
ute
7à9
Less compact, very practical:Assignment map
Number for elements for storing allshortest routes in the network:
( 1) 12 9 8 864A N N´ ´ - = × × =
302400010000
Alternative representation of paths
![Page 59: CIE4801 Transportation and spatial modelling (Uncongested ...](https://reader034.fdocuments.in/reader034/viewer/2022051600/627fa84b1a7c80231a2ed1a2/html5/thumbnails/59.jpg)
59CIE4801: (Uncongested) Assignment
How to use an assignment map
Link costs
Assignment map
For every OD a row of 0/1 indicating whichlinks are part of the route
OD fl
ows
Link flows
Rout
e co
sts
All calculations can be performed using matrix algebra
![Page 60: CIE4801 Transportation and spatial modelling (Uncongested ...](https://reader034.fdocuments.in/reader034/viewer/2022051600/627fa84b1a7c80231a2ed1a2/html5/thumbnails/60.jpg)
60CIE4801: (Uncongested) Assignment
Example1 2 3 4 5 6 7 8 9 10 11 121 1 2 1 2 5 4 3 6 1 2 1
..0 0 1 0 0 0 0 1 0 0 0 01 0 1 0 0 0 0 1 0 0 0 00 0 0 0 0 1 0 0 0 1 1 10 0 0 0 0 0 0 1 0 0 0 01 0 1 1 0 0 0 1 0 0 0 00 0 0 0 0 0 0 0 0 1 1 10 0 0 0 0 0 0 0 0 0 1 00 0 0 0 0 0 0 0 0 0 1 1..
.. ..7 1 57 2 57 3 57 4 27 5 27 6 27 8 37 9 3.. ..
X
5+1+2+1=9
X 5+5+2+2=14
![Page 61: CIE4801 Transportation and spatial modelling (Uncongested ...](https://reader034.fdocuments.in/reader034/viewer/2022051600/627fa84b1a7c80231a2ed1a2/html5/thumbnails/61.jpg)
61CIE4801: (Uncongested) Assignment
Selected link analysis
• Assignment map tells which OD-pairs use a specific link
• So you can create an OD-matrix of OD-pairs using that link and assign that to the network
• Note that using a back node representation requires extra administration in the calculations to perform a selected link analysis
![Page 62: CIE4801 Transportation and spatial modelling (Uncongested ...](https://reader034.fdocuments.in/reader034/viewer/2022051600/627fa84b1a7c80231a2ed1a2/html5/thumbnails/62.jpg)
62CIE4801: (Uncongested) Assignment
Stochastic assignment and pathrepresentation
• Iterative scheme as in Probit is tailor-made for tree search algorithms (or vice versa)
• Assignment map approach can be used as well, especially for Logit• Generate routes first,
e.g. manually or using repetitive shortest path searches while systematically eliminating or penalizing links (k-shortest path algorithm)
• Check whether routes are realistic (e.g. large or very short detours)
• Advantage of assignment map is that you can check for overlap
![Page 63: CIE4801 Transportation and spatial modelling (Uncongested ...](https://reader034.fdocuments.in/reader034/viewer/2022051600/627fa84b1a7c80231a2ed1a2/html5/thumbnails/63.jpg)
63CIE4801: (Uncongested) Assignment
Assignment map and stochasticassignment
Link costs
Assignment map
For every OD one or more rows of 0/1 indicating which links are part of the route
OD fl
ows
Rout
e co
sts
Rout
e Fl
ows
Choice model
AON or Logit(per OD)
Link flows
Probit
![Page 64: CIE4801 Transportation and spatial modelling (Uncongested ...](https://reader034.fdocuments.in/reader034/viewer/2022051600/627fa84b1a7c80231a2ed1a2/html5/thumbnails/64.jpg)
64CIE4801: (Uncongested) Assignment
Switch from link level to route level
tr1 = N (15,2)
tr 2 = N (15,2)
State 1
State 2
tr1 = 15.85
tr 2 = 14.24
tr1 = 16.78
tr 2 = 13.57
Route 1=15.83Route 2=16.04
Route 1=13.46Route 2=13.56
Differences between routes is much larger due to (implicit) different assumption for time of link 1 within a given state
This approach is therefore not correct
Thus sampling route travel times instead of link travel times
![Page 65: CIE4801 Transportation and spatial modelling (Uncongested ...](https://reader034.fdocuments.in/reader034/viewer/2022051600/627fa84b1a7c80231a2ed1a2/html5/thumbnails/65.jpg)
65CIE4801: (Uncongested) Assignment
Effect of sampling at link level versus route level
Travel timeRoute 1 Route 2
Distribution of the perception of the route travel time in case of sampling at link level
Distribution of the perception of the route travel time in case of sampling at route level
Note that compared to the previous example the average travel times of the routes differ. The Standard deviation however is similar.
![Page 66: CIE4801 Transportation and spatial modelling (Uncongested ...](https://reader034.fdocuments.in/reader034/viewer/2022051600/627fa84b1a7c80231a2ed1a2/html5/thumbnails/66.jpg)
66CIE4801: (Uncongested) Assignment
Probit and logit at route level
• For the error term at route level both extreme value distribution (=>logit) or normal distribution (=> probit) can be assumed
• In both cases the same error is made: in case of overlapping routes consistency within a network state is not guaranteed
• However, logit is much easier to compute…..
• Today, there are advanced logit models that can correct for overlap• For simple cases Nested Logit might be used
![Page 67: CIE4801 Transportation and spatial modelling (Uncongested ...](https://reader034.fdocuments.in/reader034/viewer/2022051600/627fa84b1a7c80231a2ed1a2/html5/thumbnails/67.jpg)
67CIE4801: (Uncongested) Assignment
6.Trucks
![Page 68: CIE4801 Transportation and spatial modelling (Uncongested ...](https://reader034.fdocuments.in/reader034/viewer/2022051600/627fa84b1a7c80231a2ed1a2/html5/thumbnails/68.jpg)
68CIE4801: (Uncongested) Assignment
What’s the problem(and possible solutions)
• Simply put, there are also trucks on the road……
• Add an OD-matrix of trucks to the OD-matrix of cars• Note that a truck is not a car • Use a PCU-value to correct for the size and impact:
PCU for a truck is 1.5 or 2• Assign the resulting OD-matrix of vehicles
• Assign OD-matrix of trucks separately (e.g. using different speeds) and add the flows• Use a PCU-value
![Page 69: CIE4801 Transportation and spatial modelling (Uncongested ...](https://reader034.fdocuments.in/reader034/viewer/2022051600/627fa84b1a7c80231a2ed1a2/html5/thumbnails/69.jpg)
69CIE4801: (Uncongested) Assignment
Topics Congested assignment
• What does this modelling component do? What’s its output and what’s its input? How does it fit in the framework?
• The main concepts• Wardrop’s equilibrium, speed-flow curves
• The modelling methods• Iterative scheme: Method successive averages (MSA)• Convergence criteria: Relative or Duality Gap• Mathematical programming approach• Frank-Wolfe algorithm versus MSA• Stochastic equilibrium assignment (SUE)
• Practical issues• Are these models appropriate?