You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Allow for specifying multiple initial conditions and solving in parallel using the IDAKLU solver
Motivation
Solving many simulations in parallel with different initial conditions is neccessary for implementing Unscented Kalman Filter or Particle Filter based parameter estimation methods, such as those currently being implemented in PyBOP (pybop-team/PyBOP#12). The IDAKLU solver currently has functionality for running multiple simulations with different parameters, but not for different initial conditions
Possible Implementation
How this is implemented would have to be discussed further, as currently the initial conditions are part of the model rather than the solver. Perhaps the solver could have an API for overriding the model y0 which allows for passing in either a column vector (single simulation) or a 2D array (n simulations where n is the number of columns)
Additional context
No response
The text was updated successfully, but these errors were encountered:
Description
Allow for specifying multiple initial conditions and solving in parallel using the IDAKLU solver
Motivation
Solving many simulations in parallel with different initial conditions is neccessary for implementing Unscented Kalman Filter or Particle Filter based parameter estimation methods, such as those currently being implemented in PyBOP (pybop-team/PyBOP#12). The IDAKLU solver currently has functionality for running multiple simulations with different parameters, but not for different initial conditions
Possible Implementation
How this is implemented would have to be discussed further, as currently the initial conditions are part of the model rather than the solver. Perhaps the solver could have an API for overriding the model y0 which allows for passing in either a column vector (single simulation) or a 2D array (n simulations where n is the number of columns)
Additional context
No response
The text was updated successfully, but these errors were encountered: