Order book modeling include both Limit Orders and Markets Orders arrival time simulation, one particular orientation is to perform such simulation in either continuous time, as seen more frequently in literature, in discrete time as discrete, or point process.
Problem
The atelier library does not currently support other solution rather a sampling process from a poisson distribution, which does provide a discrete event simulation with a decay factor, but there is a growing need to expand this capacity to include a clustering effect, along to a self exciting behavior.
Solution
Implement the simplest form of a Hawkes Process in order to use it for both Limit Order and Market Order arrival times simulation. The first implementation case is the univariate self-exciting with exponential decay.
There exists a gen_hawkess.rs file with a detailed example on how to use such functionality, as well its corresponding unit test in the tests/test_generators.rs file
Context
Order book modeling include both Limit Orders and Markets Orders arrival time simulation, one particular orientation is to perform such simulation in either continuous time, as seen more frequently in literature, in discrete time as discrete, or point process.
Problem
The
atelierlibrary does not currently support other solution rather a sampling process from a poisson distribution, which does provide a discrete event simulation with a decay factor, but there is a growing need to expand this capacity to include aclusteringeffect, along to aself excitingbehavior.Solution
Implement the simplest form of a Hawkes Process in order to use it for both Limit Order and Market Order arrival times simulation. The first implementation case is the univariate self-exciting with exponential decay.
$$ \begin{equation} \phi{ij}(t) = \alpha{ij} e^{-\beta_{ij} t} \end{equation} $$
Acceptance Criteria
There exists a
gen_hawkess.rsfile with a detailed example on how to use such functionality, as well its corresponding unit test in thetests/test_generators.rsfile