Timeline for On the generator of a Markov process
Current License: CC BY-SA 4.0
9 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Sep 23, 2021 at 15:50 | comment | added | Nate River | Hahaha, that description made me chuckle. Could imagine the professor saying it in real time. Thanks for the elaboration! | |
| Sep 22, 2021 at 9:12 | comment | added | Mateusz Kwaśnicki | Why? — It's a wild animal! How? — Get a domesticated one instead, like the resolvent or the semigroup. This is a partially humorous advice that I actually got from prof. Byczkowski. The main reason to avoid the use of generators is related to problems with description of its domain. The question here is a good example: the relationship between the original process and its part ("killed process") is absolutely clear, but at the level of generators it is extremely difficult to describe $L^U$ in terms of $L$, at least in the general case. | |
| Sep 22, 2021 at 5:14 | comment | added | Nate River | @MateuszKwaśnicki Could you elaborate on your advice in [*] not to use the generator if possible? Why is it undesirable to work with it, and how can one avoid it? | |
| Sep 21, 2021 at 11:42 | comment | added | sharpe | Thank you for writing a concrete example. Dealing with generators is not easy for me... | |
| Sep 21, 2021 at 10:55 | comment | added | Mateusz Kwaśnicki | This is indeed rather general, but not entirely general. I added a comment to my answer to show possible limitations. Generators are often very counter-intuitive creatures. | |
| Sep 21, 2021 at 10:54 | history | edited | Mateusz Kwaśnicki | CC BY-SA 4.0 |
added 1653 characters in body
|
| Sep 21, 2021 at 8:39 | comment | added | sharpe | Thank you for your answer! I recognized that the generator of the part process is identified with that of the original process under a relatively general setting. | |
| Sep 21, 2021 at 8:33 | vote | accept | sharpe | ||
| Sep 21, 2021 at 7:57 | history | answered | Mateusz Kwaśnicki | CC BY-SA 4.0 |