Changes between Initial Version and Version 1 of Ticket #6411, comment 8


Ignore:
Timestamp:
2021-03-11T11:05:47Z (4 years ago)
Author:
Philip Hannebohm

Legend:

Unmodified
Added
Removed
Modified

  • Ticket #6411, comment 8

    initial v1  
    22> I just claim that you can define a single *zero-crossing function* to be passed to the solver, that contains the product, and can be used to detect the event times, when one of the original zcf's changes the sign. Whether or not that corresponds to an actual event, it needs to be checked with the original condition. If there is no event, e.g. a became < 2 and b was less than 3, then you can simply continue with the next integration step.
    33
    4 Okay, I was assuming the sign information of only the zcf is used by the solver to decide if it needs to find the zero or not. But then we need to do separate checks anyway afterwards. So I don't see the benefit.
     4Okay, I was assuming the sign information of only the zcf is used by the solver to decide if it needs to find the zero or not. But then we need to do separate checks anyway afterwards. ~~So I don't see the benefit.~~
     5
     6Edit: You check one product-zcf and only if that changes sign you investigate further. Makes  sense now!