Exam four-step-procedure for travel demand estimation Flashcards
Theory questions about the four-step-procedure
In what step should a 10% growth in population be added as input? Which model parts will it affect? How is the data used?
March 21, 2022
Add as input in trip generation. Higher population will generate more trips. It will then affect all the steps after.
March 21, 2022
Where to introduce a new mode comparable to car?
March 21, 2022
Mode choice. Hard to calibrate without observations but start similar to car with parameters and attributes.
March 21, 2022
Stochastic route choice
March 20, 2023
Logit-based model where T_ij is the number of travellers in OD-pair (i,j), V_ijr is the utility for OD-pair (i,j) route r and R_pq are alla possible routes for OD-pair (p,q). Lower utility will result in less flow. The traveller will chose route based on perceived value.
March 20, 2023
Wardrop user equilibrium
March 20, 2023
Everyone travels according to their best possible choice where the system is in equilibrium when no traveller can change route to get a lower cost. All travel times (cost) will be equal for the OD-pair. h_ijr is the number of travellers choosing route r for OD-pair (i,j) and h_ijr > 0 if route is in use. The travel times for OD-pair (i.j) is given by pi_ij = sum over all links a, (delta^a)ijr ta(va) where ta(va) is the volume delay function, v is the flow and delta indicates links a used in route r for OD-pair (i,j).
March 20, 2023
Differences between stochastic route choice and Wardrop equilibrium.
March 20, 2023
There will be a difference in result since congestion is handled different in the models. Wardrop equilibrium is computationally more difficult and is based on obsereved costs while Logit models are based on what is perceived and is easier to compute once the parameters are estimated.
March 20, 2023
Information going from right to left
ExampleExam
Travel times (flows and costs data). The rational part is that it may affect mdoe choice and trip distribution. People usually prefer short trips.
ExampleExam
Shortcoming for trip generation
ExampleExam
Often based on survey data and statistical data that shows a relation between landuse and trips. For example, households and work locations. But this does not cover every relation.
ExampleExam
Shortcoming for trip distribution
ExampleExam
Shortcoming for mode choice
ExampleExam
Logit models require a large amount of data which is expensive and it is hard to collect.
ExampleExam
Shortcoming for route choice
ExampleExam
VDFs are hard to formulate, the do not recreate the reality fully since real behaviour is more dynamic than what a function can be. It also asumes static networks where nothing changes even temporarily.
ExampleExam
OD-calibration
March 19, 2021
OD-calibration adds data to the OD-matrix. The input is an OD-matrix and traffic counts. The output is a new OD-matrix. It is useful to estiamte and describe the current situation. Since it requires traffic counts, it cannot be used for forecasting becasue we cannot do traffic counts in the future. The behaviour assumption connections made in the four steps will be lost.
March 19, 2021
Input and output for route choice
March 19, 2021
Input: Description of network (the transport infrastructure), volume delay functions, demand matrix specific for a certain travel mode and time period.
Output: Flows on the links in the network and costs (travel times) associated to the flows.
March 19, 2021
What happens wehn the number of attracted trips increase?
June 8, 2023
It will change the input to the trip generation step. These trips will have to be produced somewhere so the trip production must be updated as well. Likely the numbers must be scaled to match. This will change the output which then goes in to the next step.
June 8, 2023
Improving PT with increased service
June 8, 2023
Introduce the change in the mode choice where it will have the most significant effect. Attribute values, for example waiting time, will change. If the total change in mode choice is large, this will affect the route choice and the travel times which should be fed back to the trip distribution.
June 8, 2023
Input in trip generation
March 18, 2019
Landuse, for example workplaces and residents.
March 18, 2019
