studying linear systems at the moment, and i just learned how you can combine these block diagrams to model some systems (кст, надо проверить если все линейные инвариантные к времени (LTI) системы можно так смоделировать, мне кажется да)
and if you can model them like that, it gives you another form of vector/matrix repsentation of your system
which for example you could examine, and decompose back into the blocks
these rules are not really strict, and one can figure out that they can be applied arbiturary describing the same system with different equations (or block diagrams)
да, я последние 2.5 часа разбирал 1.5 главы, решал на бумаге задачки из книги… уже 00:07
at work, i often assume that i can just use a linear regression model… just assume that there is a linear relationship; but more often than not – it’s not the case
yet we have so much mathematics for linear cases that it’s just to inconvenient to throw the linearity assumption away
so, what we can do instead is do a Taylor series expansion around a point, assuming that the pertubations caused from deviations of variables are small, such that the squared and later terms are insignificant, we can keep just the linear terms
hence allowing to use the linearity assumption
key idea to keep in mind is that this only works if the deviations are small
я тут не объяснил еще что не вокруг каждой точки так можно сделать а только вокруг equalibriums, для деталей можно прочитать скриншоты
materials like these keep me up at night