Introducing the core probability framework and discrete-element core probability model for efficient stochastic macroscopic modelling
conference paper
In this contribution the Core Probability Framework (CPF) is introduced with the
application of the Discrete-Element Core Probability Model (DE-CPM) as a new DNL
for dynamic macroscopic modelling of stochastic traffic flow. The model is
demonstrated for validation in a test case and for computational efficiency on two
simple networks. The CPF extends a base model, such as the Cell Transmission Model
(CTM), by considering each traffic variable as a discrete stochastic variable denoted
as a probability distribution of values for each traffic variable in time and space.
Traffic is propagated along a link using the base model and through a larger network
with the application of probability merging algorithms at the nodes. Due to the
incorporation of probability in the core of traffic propagation, the necessity for
multiple acts as an internalisation of the Monte Carlo routine in the CPF for fast and
efficient calculation of uncertainty. Initial tests cases show that the DE-CPM has the
potential to reduce computation time multi-tenfold compared to regular Monte Carlo
simulation. Such developments allow the application of stochastic dynamics traffic
models to be more readily applied in practice.
application of the Discrete-Element Core Probability Model (DE-CPM) as a new DNL
for dynamic macroscopic modelling of stochastic traffic flow. The model is
demonstrated for validation in a test case and for computational efficiency on two
simple networks. The CPF extends a base model, such as the Cell Transmission Model
(CTM), by considering each traffic variable as a discrete stochastic variable denoted
as a probability distribution of values for each traffic variable in time and space.
Traffic is propagated along a link using the base model and through a larger network
with the application of probability merging algorithms at the nodes. Due to the
incorporation of probability in the core of traffic propagation, the necessity for
multiple acts as an internalisation of the Monte Carlo routine in the CPF for fast and
efficient calculation of uncertainty. Initial tests cases show that the DE-CPM has the
potential to reduce computation time multi-tenfold compared to regular Monte Carlo
simulation. Such developments allow the application of stochastic dynamics traffic
models to be more readily applied in practice.
TNO Identifier
981656
Source title
DTA Symposium 2014, Salerno, Italy
Pages
1-26
Files
To receive the publication files, please send an e-mail request to TNO Repository.