Is the design of algorithms to model the physical environment on-topic? To what extent?
In particular, is What algorithm is appropriate for a thermostat controlling the heating of a room with floor heating? on-topic?
$\begingroup$
$\endgroup$
1
asked Jan 10, 2019 at 19:52
Gilles 'SO- stop being evil'ModGilles 'SO- stop being evil'
44k3 gold badges33 silver badges63 bronze badges
-
$\begingroup$ machine learning is increasingly being used for these types of problems/ applications and is a core/ critical/ emerging area of CS. $\endgroup$– vznCommented Jan 11, 2019 at 16:23
Add a comment
|
2 Answers
$\begingroup$
$\endgroup$
3
If the question is how to model the physics of the environment, I don't think that's on-topic here.
If the question is about methods for controlling the physical environment, I'm not sure. It will probably depend on whether the question is best answered from a computer science perspective or has been studied by computer scientists. In some areas (e.g., the US), control theory has been studied in electrical engineering departments rather than in computer science departments, but apparently in other areas (e.g., France), control theory is more closely associated with computer science. So, it's unclear to me where questions about PID controllers, control theory, and such should fall, but plausibly they could be on-topic here.
-
1$\begingroup$ “control theory has been studied in electrical engineering departments rather than in computer science departments" This depends on the locale. In France, control theory is more closely associated with math and CS. This is part of a more general difference between US/UK and France where in France CS historically forked from math far more often than from EE. I don't know to what extent US or France is the odd one out. $\endgroup$ Commented Jan 11, 2019 at 7:56
-
$\begingroup$ @Gilles, Ahh, I didn't know that! That sounds like an argument for control theory type questions being on-topic here, then. $\endgroup$– D.W. ModCommented Jan 11, 2019 at 8:15
-
$\begingroup$ FWIW, Germany has CS departments with both heritages. $\endgroup$– Raphael ModCommented Jan 12, 2019 at 8:54
$\begingroup$
$\endgroup$
11
First of all, I don't know if it is appropriate to post this as an answer but I need 50+ reputation points to comment and I have barely been given the rights to post an answer (thank you, Gilles).
Back to the topic on hand. The following is the same question but stripped of its context:
Given a program that outputs a value which varies with time and which can be influenced with a boolean, causing the value to rise if set to true or to fall if set to false, what algorithm would be suitable for anticipating when to switch said boolean to ensure a certain output value at a certain moment in time considering that the rate of change of the output can also vary?
It is the very same question but without 'physical world' aspect. I would assume this question is on topic, at least according to this meta answer:
Questions about designing algorithms, their correctness, or their complexity fit Computer Science best.
-
$\begingroup$ Posting an answer is fine, thanks for taking the time to enter the discussion. While you can phrase your problem purely in 'algorithmic terms' - and this might make the answer technically on-topic -, it can still not be a suitable question here, for 2 reasons. 1. With the motivation removed, the question becomes dry and difficult to understand, it is likely people will insist to clarify what you actually want instead of answering. $\endgroup$– Discrete lizard ModCommented Jan 18, 2019 at 16:43
-
$\begingroup$ 2. Nearly any question that can be formalised in mathematical term can be phrased designing an algorithm. So, it is likely that there are more experts on this problem elsewhere. In this case, that is partially true, as this question is mostly in the domain of control theory, which is familiar to some people in CS (see the discussion on the other answer) $\endgroup$– Discrete lizard ModCommented Jan 18, 2019 at 16:43
-
$\begingroup$ Thank you for the feedback. I would have asked you in a chat room but I need 19 more points to do so. Now, I don't mean to criticize, but the first thing I did after posting the question was argue as to the suitability of the question. Then, I had to wait for the question to be unblocked after its suitability was debated here. After all of this time I am still no closer to knowing where I should actually post my question, especially since there was only a proposal for CT that never went through. $\endgroup$ Commented Jan 18, 2019 at 18:45
-
$\begingroup$ As an advice, if you will, would it not be easier and a better experience for new users such as me if moderators would simply be able to move the question to the appropriate site? $\endgroup$ Commented Jan 18, 2019 at 18:53
-
$\begingroup$ I agree that we haven't been very helpful so far, my apologies. The main problem here is that we do not know if there is a better site for this question is either. This means we have focused on the easier question, which is whether this question is should be on this site or not. However, there are some sites where this might work. I'll go ask around. $\endgroup$– Discrete lizard ModCommented Jan 19, 2019 at 11:00
-
$\begingroup$ As for your advice, that is standard procedure (IIRC), if 1. The question is clearly off-topic and 2. the moderators know at which site the question is on-topic. This meta thread was created because we do not know 1 yet. Note that part 1 is a requirement mainly because we prefer not to move questions around unless needed, since some questions can be on-topic here and reasonably answered even though it would fit 'better' on other sites. (but I think that your question would do better elsewhere) $\endgroup$– Discrete lizard ModCommented Jan 19, 2019 at 11:04
-
$\begingroup$ While there is no site for control theory, control theory is an important tag on Electrical Engineering. I've asked in their chat whether your question is suitable for their site. $\endgroup$– Discrete lizard ModCommented Jan 19, 2019 at 11:22
-
$\begingroup$ Thanks for the help. So should I post it on the home improvement channel or will you move it if / when suitable? $\endgroup$ Commented Jan 20, 2019 at 10:45
-
$\begingroup$ Since you have already an answer to your question, it is best to request migration, since then that answer moves together with your question. You should flag your question for moderator attention and explain that you want it to be migrated to the home improvement site. $\endgroup$– Discrete lizard ModCommented Jan 20, 2019 at 10:50
-
$\begingroup$ Ok, but does everyone agree that is the appropriate site for the question? $\endgroup$ Commented Jan 20, 2019 at 11:09
-
$\begingroup$ Well, I'm not sure about 'everyone', but I think it wouldn't hurt to try. There seems to be consensus here that Computer Science is not the right place for this question. If a question gets migrated to another site, then the people on that site can still decline if they don't want it. So, as long as you agree that the question would fit better there, you should just ask. $\endgroup$– Discrete lizard ModCommented Jan 20, 2019 at 11:17