Changes between Initial Version and Version 1 of Ticket #5836, comment 7


Ignore:
Timestamp:
2020-06-18T22:33:53Z (5 years ago)
Author:
Francesco Casella

Legend:

Unmodified
Added
Removed
Modified
  • Ticket #5836, comment 7

    initial v1  
    4040The rationale of this restriction is by no means to make those equations assignments with fixed causality, or to make them ''de facto'' algorithms. This is not written anywhere, and I understand it is also not intented.
    4141
    42 The rationale of this structural restriction is instead that when-equations are activated only at event instants, when they are (conceptually) solved together with all the other non-when-equations that are always active (single assignment rule). Otherwise, if the when-equation is not activated, ''its corresponding discrete variable is kept constant''. In practice, you have many blocks in the BLT, some of them matched to discrete variables, some to continuous variables, and you only execute the blocks that are active at the current step; if you end up with blocks matched with mixed continuous and discrete equations, you give up. At least that's what Dymola does, and it's fine, because most of the time the mixed system is the result of having incorrectly omitted some {{{pre()}}} operators.
     42The rationale of this structural restriction is instead that when-equations are activated only at event instants, when they are (conceptually) solved together with all the other non-when-equations that are always active (single assignment rule). Otherwise, if a when-equation is not activated, ''its corresponding discrete variable is kept constant''. In practice, you have many blocks in the BLT, some of them matched to discrete variables, some to continuous variables, and you only execute the blocks that are active at the current step; if you end up with blocks matched with mixed continuous and discrete equations, you give up. At least that's what Dymola does, and it's fine, because most of the time the mixed system is the result of having incorrectly omitted some {{{pre()}}} operators.
    4343
    44 The question now is is: what is the corresponding variable? This structural restriction makes it clear: it' the one showing up in the left-hand side.
     44The question now is is: what is the corresponding variable? This structural restriction makes it clear: it's the one showing up in the left-hand side of the when-equation.
    4545
    4646Alles klar?