residue / congruence class
The congruence class of is an equivalence class.
is the set of all rest classes (factor group).
E.g. for :
→ the set of even numbers.
→ the set of odd numbers.
So basically, is the set of all numbers that are divisible by , and is the set of all numbers that have a remainder of 1 when divided by etc.
In words: two rest classes are either identical or have nothing in common (“disjoint”; no element of is in ; = empty set).
Proof:
Assumption: (see the definition of rest class above, then the implication is apparent)
(just rewrite with definition of divisibility)
WLOG: (the underlined things lead to the implication, the boxed thing is a contradiction)
Operations in Let We can translate these two congruences to: → We can add/sub and the congruence stays: → Same for multiplication: ( differs from by a multiple of → definition of congruence)
Some congruences:
, (odd, i.e. in )
(even, i.e. in )
→
(still odd)
(still even)
Note: , i.e. the congruence is still true.
Here’s why we can do this:
Link to original… generalizing modulo arithmetic …
Different representations of the same coset/element in the factor group.
For the abelian group and , its cosets area normal subgroup .
Looking at the concrete case of , we have the following congruence classes:
(note: is the neutral element of the factor group)
We can do arithmetic with these classes, e.g.:
This is the same as the addition in , adding the elements and taking the remainder modulo 4.
Same as above, but a different representation:
The reason we can do this is because .
For any group with a normal subgroup and the coses (left and right coset equiv).
Now (like above) we define the same operation on the cosets as on the group, i.e.
We can now take this concept of modulo arithmetic and generalize it it to any group with a normal subgroup.
,
Proof:
( are representatives of the same coset)
(since and being a subgroup and hence closed).
From the operation rules earlier:
(one can proove by induction)
(n as a decimal number)
In words: a number is divisible by 3 if the sum of its digits is divisible by 3.
The same works for too.
→ (we can directly operate on the rest classes)
E.g. , (even + even = even, odd + even = odd)
Operation table (always read row column) for addition:
Operation table for multiplication:
Properties of
commutative, associativity, distributive property hold for the operations in .
There is also a neutral element: for addition, for multiplication.
(; elements are invertible – already present in the operation table, just represented without a minus.)
wrong (not all elements are invertible for multiplication; no division by zero)
Example: Multiplication table for : (see that is not invertible)
In words: is invertible in if and only if and are coprime.
Proof:
“” (assumption; going left to right)
( and only differ by a multiple of … )
“”
E.g. Find
→
Division and divisibility
Since and are coprime, we can divide by , aka multiply by :
If we have:
We can’t divide by , because the modul contains it, they are not coprime, there is no inverse.
However, if
In words: If there is a common divisor on the left side, the right side, and in the modulo, we can cancel it out / divide by it.
In words: All rest classes with a prime modulus (except for ) have a multiplicative inverse.
ISBN
(since there are no , it’s a concatenation of digits, “multiplikation im Monoid der Strings”)
Prüfziffer (check digit):
( is treated as )Theorem: Single errors and transpositions of two digits can be detected (but not corrected).
Proof for single errors:
Let’s assume there is some error at position :
,
,
We check whether the sum is congruent modulo for the original and the changed number:
Since all possible are coprime and thus invertible modulo , we can divide by :
Since are less than the modulo, they have to be equal, else the congruence would not hold.Proof for single transpositions:
(transposition of two digits):
,
(same reason as before)
→ The code can only be correct if the sum is congruent modulo .