csa

docs/quetzal_library/transport_modeling_intro.md

@@ -34,18 +34,23 @@ Quetzal was inspired by the classical 4 step transport modeling and provides alg
 
 # Assignment
 ## Optimal Strategies
-
 - *Optimal Strategies* can be established toward the destinations
 - It a set of relevant line to board at each station and associated alighting stations
 - The path is not known beforehand but is built along the journey of the traveler
 - The expected time of arrival is better than the better sequence of lines yielded by the shortest path algorithm run in a frequency-based graph
 - The assignment on each link is derived from the deterministic strategies
 ![classical four step structure](/transport_modeling/OS_map.png)
-On this complex example in Dire Dawa (Ethiopia), users start from Misrak Buna in the east and go to Kebele 2 in the west
-Then they wait in Lega Hare, for the 10 or 10bis to stop and board the first one.
-Then, those who board line 10 alight in Legehar and the users of line 10 bis alight in Ashawa.
-Both line 1 and line 22bis stop at Ashawa (first stop) and Legehar. Our users will board the first of these line to stop at their station. 
-Once they are in line 1 or 22bis, they will alight in Bridge to board the line 18 and alight at Jerba Police then walk to their final destination.
+On this complex example in Dire Dawa (Ethiopia):
+- users start from Misrak Buna in the east and go to Kebele 2 in the west
+- Then they wait in Lega Hare, for the 10 or 10bis to stop and board the first one.
+- Then, those who board line 10 alight in Legehar and the users of line 10 bis alight in Ashawa.
+- Both line 1 and line 22bis stop at Ashawa (first stop) and Legehar. Our users will board the first of these line to stop at their station. 
+- Once they are in line 1 or 22bis, they will alight in Bridge to board the line 18 and alight at Jerba Police then walk to their final destination.
 ![classical four step structure](/transport_modeling/OS_flowchart.png)
-
-
+## Connection Scan Algorithm
+- *Connection Scan Algorithm* is a timetable-based pathfinder 
+- Provides a departure time vs arrival time pareto-optimal (non dominated) set of paths
+- A path dominates another if it has both a later departure time and an earlier arrival
+- Can be run in public transport subgraphs (pruned graphs) to yield non pareto optimal itineraries 
+- The route choice can be integrated with the mode choice and the time-of-day choice in a nested logit by discretizing the origin-destination matrix in origin-destination-wished departure time matrix
+![classical four step structure](/transport_modeling/CSA.png)