2.5.

Posets
In this module we introduce the concepts of Posets and totally ordered set. We
illustrate the method of drawing a Hasse diagram of a finite poset and study
Lexicographic ordering. At the end we discuss the technique of topological
sorting.

We often use relations to order some or all of the elements of sets. For example
(i) we order words using the relation containing pairs of words (𝑥, 𝑦), where the
word 𝑥 comes before the word 𝑦 in the dictionary, (ii) we schedule projects using
the relation consisting of pairs (𝑥, 𝑦), where 𝑥, 𝑦 are tasks in a project such that
the task 𝑥 must be completed before the task 𝑦 begins, (iii) we order the
elements of a set of real numbers using the relation containing the pairs (𝑥, 𝑦)
where 𝑥 is less than 𝑦. If we add all the pairs of the form (𝑥, 𝑥) to these relations,
we obtain the relation that is reflexive, antisymmetric and transitive. These are
the properties that characterize relations used to order the elements.

Partial order and Poset

A relation 𝑅 on a set 𝐴 is called a partial ordering or a partial order if it is

reflexive, antisymmetric and transitive. A set 𝐴 together with a partial order 𝑅
defined on it, denoted by (𝑨, 𝑹) is called a Partially ordered set or Poset in short.

Examples

i) The relation “less than or equal to ≤ ” and “greater than or equal to ≥ “

are partial orders on any nonempty subset 𝐴 of 𝑹. Therefore, (𝐴, ≤) and
(𝐴, ≥) are posets, where 𝐴 is a nonempty subset of 𝑹.
ii) The divisibility relation (i.e., divides (|)) is a partial order on any nonempty
subset 𝐴 on 𝑵. Therefore, (𝐴, |) is a poset, where 𝐴 is a nonempty subset of
𝑵.
iii) Let 𝐴 be a set and 𝑃(𝐴) be its power set. The inclusion relation ⊆ is a partial
order on 𝑃(𝐴). Therefore, 𝑃 𝐴 , ⊆ is poset, where 𝐴 is a set.
Notation: Customarily a partial order is denoted by ≼.

This notation is used because the relation ≤ is the most familiar example of a
partial order on 𝑹 and the symbol ≼ is similar to ≤.

The notation 𝑥 ≺ 𝑦 denotes that 𝑥 ≼ 𝑦 and 𝑥 ≠ 𝑦.

Comparability in a poset

Let (𝐴, ≼) be a poset. Two elements 𝑎, 𝑏 ∈ 𝐴 are said to be comparable if either

𝒂 ≼ 𝒃 or 𝒃 ≼ 𝒂. The elements 𝑎, 𝑏 ∈ 𝐴 are incomparable if neither 𝒂 ≼ 𝒃 nor
𝒃 ≼ 𝒂.

Example 1: In the poset 𝑃 𝐴 , ⊆ , where 𝐴 = 1, 2, 3 ; 1 and 1, 3 are

comparable (because 1 ⊆ 1, 3 ) and 1, 2 , 1, 3 are incomparable because
neither set is a subset of the other.

The adjective partial in the poset means that there may be pairs of elements
which are incomparable. If every pair of elements in a poset are comparable then
the ordering relation is called a total order.

Totally ordered set: If (𝐴, ≼) be a poset and every two elements of 𝐴 are
comparable then (𝐴, ≼) is called a totally ordered set or linearly ordered set or
chain and ≼ is called a total order or a linear order.

Example 2: Let 𝐴 be any nonempty subset of 𝑹. The poset (𝐴, ≤) is a totally

ordered set because for any 𝑎, 𝑏 ∈ 𝐴, we have either 𝑎 ≤ 𝑏 or 𝑏 ≤ 𝑎.

Example 3: Let 𝐴 = 1, 2, 3, 4 . The poset (𝐴, |) is not a chain because 𝐴 contains

elements that are incomparable such as 2 and 3.

Consider the totally ordered set (𝑵, ≤) and 𝑵 × 𝑵. Define a relation ≼ on 𝑵 × 𝑵

as follows. For any 𝑥1 , 𝑦1 , 𝑥2 , 𝑦2 ∈ 𝑵 × 𝑵 , define
𝒙𝟏 , 𝒚𝟏 ≼ 𝒙𝟐 , 𝒚𝟐 either if 𝒙𝟏 < 𝒙𝟐 or if both 𝒙𝟏 = 𝒙𝟐 and 𝒚𝟏 ≤ 𝒚𝟐
Example 4: 𝑵 × 𝑵, ≼ is a totally ordered set (See P1 for solution)

Lexicographic order

The words in a dictionary are listed in alphabetic or lexicographic order which is

based on the ordering of the letters in the alphabet. This is a special case of an
ordering of strings on a set constructed from a partial ordering on the set. The
following is such a construction in any poset.

Let (𝐴1 , ≼1 ) and (𝐴2 , ≼2 ) be two Posets. Consider the Cartesian product 𝐴1 × 𝐴2.
The lexicographic ordering ≼ on 𝐴1 × 𝐴2 is defined by

𝒙𝟏 , 𝒚𝟏 ≼ 𝒙𝟐 , 𝒚𝟐 ⟺ 𝒙𝟏 ≺𝟏 𝒙𝟐 ∨ 𝒙𝟏 = 𝒙𝟐 ∧ 𝒚𝟏 ≼𝟐 𝒚𝟐

It may be verified that ≼ is a partial order. Therefore, 𝐴1 × 𝐴2 , ≼ is a poset.

If (𝑨𝟏 , ≼𝟏 ) and (𝑨𝟐 , ≼𝟐 ) be totally ordered sets then so is 𝑨𝟏 × 𝑨𝟐 , ≼

Example 5: In the totally ordered set 𝑵 × 𝑵, ≼ , where ≼ is the lexicographic

ordering constructed from the usual ≤ relation on 𝑵,

2, 2 ≼ 2, 1 , 3, 1 ≼ 1, 5 , 2, 2 ≼ 2, 2 , 3, 2 ≼ 1, 1 , 4, 9 ≼ 4, 11 , …

Generalization of Lexicographic ordering

Let (𝐴, ≼1 ) be a totally ordered set. Let 𝑛 be a given natural number and
𝑛

𝑃 = 𝐴 ∪ 𝐴2 ∪ 𝐴3 ∪ … . 𝑈𝐴𝑛 = 𝐴𝑖
𝑖=1

That is, 𝑃 consists of strings of elements of 𝐴 of length less than or equal to 𝑛.

A string of length 𝑝 may be considered as a 𝑝-tuple .
Define a total ordering ≼ on 𝑃, called lexicographic ordering, as follows:

Let 𝑎1 , 𝑎2 , … , 𝑎𝑝 and (𝑏1 , 𝑏2 , … , 𝑏𝑝 , … , 𝑏𝑞 ) with 𝑝 ≤ 𝑞 be any two elements of

𝑃. Now,

(𝑎1 , 𝑎2 , … , 𝑎𝑝 ) ≼ (𝑏1 , 𝑏2 , … , 𝑏𝑝 , … , 𝑏𝑞 )
if any of the following hold:

i) 𝑎1 , 𝑎2 , … . , 𝑎𝑝 = (𝑏1 , 𝑏2 , … , 𝑏𝑝 )
ii) 𝑎1 ≠ 𝑏1 and 𝑎1 ≺1 𝑏1 in (𝐴 ≼1 )
iii) 𝑎𝑖 = 𝑏𝑖 , 𝑖 = 1,2, … , 𝑘 𝑘 < 𝑝 and 𝑎𝑖+1 ≠ 𝑏𝑖+1 and 𝑎𝑖+1 ≺1 𝑏𝑖+1 in (𝐴, ≼1 )

If none of these conditions are satisfied, then

(𝑏1 , 𝑏2 , … , 𝑏𝑞 ) ≼ (𝑎1 , 𝑎2 , … , 𝑎𝑝 )

Example 6: Let 𝐴 = 𝑎, 𝑏, 𝑐, … , 𝑧 be the lower case English alphabet and there is

a linear ordering on 𝐴 denoted by ≤, where 𝑎 ≤ 𝑏 ≤ 𝑐 ≤ ⋯ ≤ 𝑧. Let

𝑃 = 𝐴 ∪ 𝐴2 ∪ 𝐴3 ∪ 𝐴4 ∪ 𝐴5

That is ,𝑃 consists of all words or strings of 5 or fewer than 5 letters from 𝐴. The
lexicographic ordering in 𝑃 is same as that used in dictionaries

For example, ante ≼ axe (by ii), zebra ≼ zero (Here “zero” and “zebra” are
compared and the conditions (i), (ii) and (iii) are not satisfied) and
zerm ≼ zero (by (iii))

Note: Instead of using ≼ to denote lexicographic ordering, it is customary to use

the terminology such as lexically less than or equal to or lexically greater than

Application of lexicographic ordering

Lexicographic ordering is used in sorting character data on a computer

Hasse diagrams
We can simplify the diagraphs of a finite poset by omitting many of its edges. For
instance, since a partial order is reflexive, each vertex has a loop, which we can
delete. In addition, drop all edges implied by transitivity (For example if the
digraph of a poset contains edges 𝑎, 𝑏 and (𝑏, 𝑐) , then it has the edge 𝑎, 𝑐 ,
which we can omit. Finally, draw the remaining edges upward and drop all
arrows. The resulting diagram is called the Hasse diagram, named after the
twentieth – century German mathematician Helmut Hasse (1898 − 1979) .
Example 7: Draw the Hasse diagram for the poset (𝑨, |), where
𝑨 = 𝟏, 𝟐, 𝟑, 𝟔, 𝟖, 𝟏𝟐 and | denotes the divisibility relation.

Solution: The ordered pairs in the partial order are

1,1 , 1,2 , 1,3 , 1,6 , 1,8 , 1,12 , 2,2 , 2,6 , 2,8 , 2,12 , 3,3 , 3,6 , (3,12)

6,6 , 6,12 , 8,8 , (12,12)

The diagraph of the poset is

Step 1: Delete the loop at each vertex

Step 2: Delete all the edges implied by transitivity. These are
1,6 , 1,8 , 1,12 , 2,12 , (3,12)

Step 3: Omit all arrows and arrange all edges pointing upward to obtain the Hasse
diagram.

Example 8: Draw Hasse diagram of the poset (𝑷 𝑨 , ⊆) , where 𝑨 = 𝒂, 𝒃, 𝒄

Solution: We have 𝐴 = 𝑎, 𝑏, 𝑐 and

𝑃 𝐴 = ∅, 𝑎 , 𝑏 , 𝑐 , 𝑎, 𝑏 , 𝑏, 𝑐 , 𝑐, 𝑎 , 𝑎, 𝑏, 𝑐

Following the steps 1-3, produces the Hasse diagram of the poset (𝑷 𝑨 , ⊆)
Example 9: Draw the Hasse diagram of the poset 𝑨, ), where
𝑨 = 𝟐, 𝟑, 𝟔, 𝟏𝟐, 𝟐𝟒, 𝟑𝟔

Solution: We have 𝐴 = 2,3,6,12,24,36 . Following the steps 1-3, produces the

Hasse diagram of 𝑨, ),
Extremal elements: An element 𝑎 in a poset 𝐴, ≼ is a maximal element, if there
is no element 𝑏 in 𝐴 such that 𝑎 ≺ 𝑏. An element 𝑎 in a poset 𝐴, ≼ is a minimal
element, if there is no element 𝑏 in 𝐴 such that 𝑏 ≺ 𝑎.

The maximal and minimal elements in a finite poset are the top and bottom
elements in its Hasse diagram.

The poset 𝐴, | , where 𝐴 = 1, 2, 3, 6, 8, 12 has two maximal elements 8 and 12

and one minimal element 1.
The poset 𝐴, | where 𝐴 = 2, 3, 6, 12, 24, 36 has two maximal elements 24 and
36 and two minimal elements 2 and 3.

Note:

(i) A poset may have more than one maximal element and more than one
minimal element.
(ii) A poset need not have any maximal or minimal elements. For example the
poset 𝒁, ≤ has no maximal or minimal elements.
(iii) A poset may have a maximal element but no minimal elements or a
minimal element but no maximal elements. For example 𝒁−, ≤ has a
maximal element but no minimal elements, where as the poset 𝑵, ≤ has
a minimal element but no maximal elements.

Although an arbitrary poset need not have a minimal element, every non empty
finite poset has at least one minimal element. We state this result as a lemma.

Lemma 1: Every finite nonempty poset 𝑨, ≼ has at least one minimal

element.

Two special extremal elements are the greatest and the least elements.

Greatest and least elements: Let 𝐴, ≼ be a poset. If there exists an element

𝑎 𝜖 𝐴 such that 𝑎 ≼ 𝑥, for all 𝑥 ∈ 𝐴, then 𝑎 is called the least element of 𝐴. If
these exists an element 𝑏 ∈ 𝐴 such that 𝑥 ≼ 𝑏 for all 𝑥 ∈ 𝐴, then 𝑏 is called the
greatest element of 𝐴.
Note:

i. The least element (greatest element) if exists is unique and they are the
bottom most (topmost) elements in the Hasse diagram of a finite poset.
ii. The least element and the greatest element of a poset are usually denoted
by 0 and 1 respectively

The poset 𝐴, | where 𝐴 = 1, 2, 3, 6, 8,12 has no greatest element, but has the
least element 1.

The poset 𝐴, | where 𝐴 = 2,3, 6, 12, 24, 36 has no least element and has no
greatest element.

Topological Sorting
Suppose that a project is made up of 𝑛 different tasks say 𝑡1 , 𝑡2 , 𝑡3 , … , 𝑡𝑛 .Some
tasks can be completed only after others have been finished. To find an order 𝑅
for these tasks, we set up a partial order on the set of tasks so that 𝑡𝑖 𝑅 𝑡𝑗 , 𝑖 ≠ 𝑗,
iff the task 𝑡𝑗 cannot be started until the task 𝑡𝑖 has been completed. To produce
a schedule for the project, we need to produce an order for all 𝑛 tasks that is
compatible with this partial order.

A total ordering ≼ is said to be compatible with the partial ordering 𝑅, if

𝒂 ≼ 𝒃 whenever 𝒂𝑹𝒃

Constructing a compatible total ordering from a given partial ordering is called

topological sorting.

The following is the procedure of a topological sorting and it works for any finite
nonempty poset.
To define a total ordering on the poset 𝐴, 𝑅 where 𝐴 has 𝑛 elements , first
choose a minimal element, say 𝑎1 (such an element exists by Lemma 1). Now
note that 𝐴 − 𝑎1 , 𝑅 is also a poset (where 𝑅 is the restriction on 𝐴 − 𝑎1 ). If
it is nonempty choose a minimal element, say 𝑎2 ,of this poset and we have
𝑎1 ≺ 𝑎2 .Remove 𝑎2 from 𝐴 − 𝑎1 , if 𝐴 − 𝑎1 𝑎2 is nonempty continue the
procedure. Because 𝐴 is finite this process must terminate.
The end product is a sequence of elements 𝑎1 , 𝑎2 , … , 𝑎𝑛 and the desired total
ordering ≼ is defined by

𝒂𝟏 ≺ 𝒂𝟐 ≺ 𝒂𝟑 ≺ 𝒂𝟒 ≺ ⋯ ≺ 𝒂𝒏

This total ordering ≼ is compatible with the original partial ordering 𝑅.

Application: Topological sorting has an application to the scheduling of projects.

Example 10: Topologically sort the elements of the poset 𝑨, | , where

𝑨 = 𝟏, 𝟐, 𝟒, 𝟓, 𝟏𝟐, 𝟐𝟎
Find a compatible total ordering for the poset 𝑨, |

Solution: The Hasse diagram of the poset 𝑨, | is

Step 1: Choose a minimal element. This must be 1 (because 1 is the only minimal
element.

Step -2: Extract 1 and obtain 𝐴 − 1 = 2, 4, 5, 12, 20 . Now, There are two
minimal elements, namely 2 and 5, select 5 .

Repeat step – 2: Extract 5 and obtain 𝐴 − 1, 5 = 2,4, 12, 20 .The minimal

element at this stage is 2.

Repeat Step – 2: Extract 2 and obtain 𝐴 − 1, 5, 2 = 4, 12, 20 . The minimal

element at this stage is 4.

Repeat Step – 2: Extract 4 and obtain 𝐴 − 1, 5, 2, 4 = 12, 20 . Either of the two

can be a minimal element, select 20.

Repeat Step – 2: Extract 20 and obtain 𝐴 − 1, 5, 2, 4, 20 = 12 Now, it is the

last element left.
This produces a total ordering 1 ≺ 5 ≺ 2 ≺ 4 ≺ 20 ≺ 12.

Note: It is one possible order for the tasks.

The steps used by this sorting are illustrated in the figure.

Note:

Computer scientists use the terminology Topological sorting and mathematician

use the terminology Linearization of partial ordering for the same thing. In
mathematics topology deals with the geometry. In computer science, a topology
is any arrangement of objects that can be connected with edges.