As far as I know, given $U=\{1,2,3,4,5,6\},A=\{1,2,3\}$ the notation for its set complement is $A^C = \{4,5,6\}$

Is there any sort of notation for a conditional set complement? For example, lets say I had a true/false variable $x_1$ who determines in an equation if $A$ should be itself or its complement. I think I could do a piece-wise function like so:

$$B=\begin{cases} A& \text{if $x_1$ is true},\\ A^C& \text{if $x_1$ is false}. \end{cases}$$

but I actually have many conditionally complemented sets that I am using. If I use a single piece-wise function, that would be $2^{n}$ cases, or I could use set operators between $n$ different piece-wise functions, but that seems very verbose. Thanks!

  • $\begingroup$ What exactly are you trying to do? For some applications of this, the symmetric difference might do what you want. $\endgroup$ – Milo Brandt Jun 12 '16 at 20:14
  • $\begingroup$ I have "tests" that I can run on an item. Each test will tell me if the item belongs to a set of elements or not. So I can take item $i$, run test $1$ and that'll return my $x_1$. I can then run further tests to determine if $i$ is in another set and take the intersections of the results of these tests. The goal is to find what element $i$ actually is, in the least number of tests since the tests are expensive. However, I don't know of any notation that can represent what I'm trying to achieve. Does that help clarify, or just muddle the problem? $\endgroup$ – Jonathan Gawrych Jun 12 '16 at 20:36
  • 1
    $\begingroup$ I see. If it were me, I might do something like say $x_{j,+}$ is the set of $i$ that would pass test $j$ and say $x_{j,-}$ is the set that would fail and demand $x_{j,+}=x_{j,-}^C$. You could also treat every test as a function $f_j$ from the input set to $\{0,1\}$ and treat the preimage $f_j^{-1}(0)$ and $f_j^{-1}(1)$ - i.e. the inputs leading to either of these results. Neither of these is specifically does what you want, but I think they're reasonable ways to express the problem. $\endgroup$ – Milo Brandt Jun 12 '16 at 20:56

Could you say:

For $i=1,\dots,n$ let

$$B_i=\begin{cases} A_i& \text{if $x_i$ is true,}\\ A_i^C& \text{if $x_i$ is false.} \end{cases}$$

Now let $B=B_1\cap\dots\cap B_n$.

Or something to that effect?

  • $\begingroup$ That seems the most reasonable and least complicated. It took me way too long to even remember "set compliment" so I wanted to make sure I wasn't overlooking any standard notation. Especially if it might have lead me to more information. $\endgroup$ – Jonathan Gawrych Jun 12 '16 at 23:31

Usually, the symbol backslash is used to denote relative complements, and if that's what you mean then what I mean is if $A$ and $B$ are sets we would write $A$ \ $B$={$x$:$x$ is in $A$ but is not in $B$}.

Never mind that's probably not what you mean.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.