There is no function that can do that simultaneously, at least not the first time.
For any given row and column exchange pattern, it is possible to pre-process the pattern so that doing the same exchange for multiple different arrays would take place simultaneously for that one matrix. Something like
temp = preprocess_exchange(rows_to_exchange, columns_to_exchange);
newA = A(temp);
newB = B(temp);
where preprocess_exchange is a function that would need to be written.