Section 3 The Summation Notation

You will frequently deal with complicated expressions involving a large number of
additions. Often, these expressions are simplified using the ‘summation’ notation.
Many students find difficulty in manipulating such expressions. The purpose of this
section is to introduce the notation to you, and to get you comfortable with it.

3.1 Definition and Rules for Sums

The uppercase sigma “  ” is used to denote summation. For an arbitrary set of

numbers {x1 , x2 ,...., xn } define
n
 xi  x1  x2  ...  xn .
i 1

Example 3.1.1 The average of a set of numbers x1 , x2 ,..., xn can be written

1 n
x   xi .
n i 1

Example 3.1.2 Write the sum 4 + 8 + 12 + 16 + 20 + 24 in summation notation:

Ans:  i 1 4i .
6

Example 3.1.3 Suppose the following payments are to be made: a1 in the first
period, a2 in the second period, and so on until an in the n th period. At a fixed
interest rate of r per period, the present value of the payments is
n
a1 a2 an a
1 r

(1  r ) 2
 ... 
(1  r ) n
  (1  ir )i .
i 1

In Example 3.1.1, the “index of summation” i enter as subscripts, but otherwise do

not enter into the computation of the terms of the summation. In Example 3.1.2, the
index is actually part of the computation of the terms. In Example 3.1.3, the index is
used both ways.

Matrix Algebra Notes 3-1

Anthony Tay
Example 3.1.4 Economists often use an aggregate price index to track the overall
price level in an economy relative to some base year. This is usually done by tracking
a weighted average of prices of a certain set of commodities. Let
i  1,..., n represent n commodities
q 0i be the quantity of good i purchased in period 0 (the base year)
p 0i be the price of good i in period 0
qti be the quantity of good i purchased in period t
pti be the price of good i in period t

 i 1 p t i q 0i
n

The Laspeyres Price Index is

 i 1 p 0i q 0i
n

 i 1 p t i q t i
n

The Paasche Price Index is

 i 1 p 0i q t i
n

Expressions using summation notation are not unique; more than one expression can
be used to represent a given sum.

Example 3.1.5 Write 1 – 1/3 + 1/5 – 1/7 + 1/9 – 1/11 in summation notation:

1 1
 i 1 (1)i 1 2i  1 . Alternate answer:  i 0 (1)i 2i  1
6 5
Ans:

3.2 Rules for Working with the Summation Notation

The summation notation greatly simplifies notation (once you get used to it), but this
is only helpful then you know how to manipulate expressions written in it. There are
only two rules to learn
n n n n n
(i)  (ai  bi )   ai   bi , (ii)  cai  c ai , where c is a constant.
i 1 i 1 i 1 i 1 i 1

 i 1 c
n
Example 3.2.1  nc
n
1 n
Example 3.2.2  ( xi  x )  0 , where x   xi .
n i 1
i 1

n n n
Proof:  ( xi  x )   xi   x  nx  nx  0
i 1 i 1 i 1

Matrix Algebra Notes 3-2

Anthony Tay
 n n n
Example 3.2.3  ( xi  x )( yi  y )   ( xi  x ) yi   ( yi  y ) xi ,
i 1 i 1 i 1
n n
1 1
where x  
n i 1
xi , and y   yi
n i 1
n n n
Proof We have  ( xi  x )( yi  y )   ( xi  x ) yi   ( xi  x ) y
i 1 i 1 i 1

n n
But the second term is zero:  ( xi  x ) y  y  ( xi  x )  0 ,
i 1 i 1

which gives the desired result.

n n
The proof of  ( xi  x )( yi  y )   ( yi  y ) xi is similar.
i 1 i 1

3.3 Some useful formulas involving summations

For every integer n,

n
n(n  1)
 i  1  2  3  ...  n 
2
i 1

n
n(n  1)(2n  1)
 i 2  12  22  32  ...  n2 
6
i 1

2
n  n 
 i3  13  23  33  ...  n3    i 
i 1  i 1 

Arithmetic Series
n 1
 i 0 (a  id )  i 1 (a  (i  1)d )
n

 a  (a  d )  (a  2d )  ...  (a  (n  1)d )
n(n  1)d
 na 
2
where a and d are real numbers.

Geometric Series
n 1 (1  r n )
 i 0 ar i   i 1 ar i 1  a  ar  ar 2  ...  ar n 1  a
n
,
(1  r )
where a and r are real numbers.

Matrix Algebra Notes 3-3

Anthony Tay
3.4 Double summations

Suppose we have a rectangular array of numbers

a11 a12  a1n
a21 a22  a2 n
   
am1 am 2  amn

Let the total sum of these numbers be S. To get S we can first add up the rows and
then add the results,
n n n m  n 
i.e., S   a1 j   a2 j     am j     aij  ,
j 1 j 1 j 1 i 1  j 1 
or first add up the columns,
m m m n  m 
i.e., S   ai 1   ai 2     ai n     aij  .
i 1 i 1 i 1 j 1  i 1 

The parenthesis makes it clear which summation is to be done first, but it is

conventional to leave out the parenthesis and write
m n n m
 aij or  aij
i 1 j 1 j 1 i 1

with the understanding that the summations are carried out from right to left, i.e.,
from the inner summation to the outer.

m n
Example 3.4.1 Expand  i j 2 .
i 1 j 1

m n m n
One way to do this is  i j 2   i j 2  (1  2   m)(12  22    n 2 )
i 1 j 1 i 1 j 1

A more explicit argument…

m n m  n 2 m  n 2  m  n 2 
 i j 2     i j         i    j 
 i j  
i 1 j 1 i 1  j 1  i 1  j 1   i 1   j 1 

 (1  2    m)(12  22    n 2 )

Matrix Algebra Notes 3-4

Anthony Tay
In the examples so far, we could interchange the summation signs, i.e.,
m n n m
 aij =  aij .
i 1 j 1 j 1 i 1

We can do this only if the limits of the outer summation do not depend on the limit of
any of the inner summations.

Example 3.4.2 Suppose we have a triangular array of numbers to be expressed in

summation notation.
a11
a21 a22
  
am1 am 2  a mm

We can write  i 1  j 1 aij . We cannot interchange the summation symbols because

m i

the upper limit in the inner summation depends on the index of the outer summation.

Example 3.4.3 Write the sum of the following triangular array using summation
notation.
a11
a21 a22
  
an,1 an,2  an,n
  
am,1 am,2  am,n
where n  m .
Solution: Let ai , j be the typical element in the sum. Then the first column has
j  1 , and i running from 1 to m ; the second column has j  2 , and i running from
2 to m . In general we have j running from 1 to n , for the j th column, i running
from j to m . Thus the sum is  j 1  i  j aij .
n m

Matrix Algebra Notes 3-5

Anthony Tay
Exercises

1. Write out in full, then evaluate or simplify

   
4 3 4 10
a. i 1
2i b. i 0
ixi c. i 1
(i  1) xi 1 d. i 1
2

   
10 10 i i
e. i 1
(2i  1) f. i 1
(1)i g. j 1
j h. j 1
i

2. Write in summation notation

a. 1  3  5  7  9

m  m  m  m 1  m  m
b. a m    a m 1b    a m  2b 2  ...    ab    b
1 2  m  1  m

c. 2 + 3/2 + 4/3 + 5/4 + 6/5

d. ai 1b1 j  ai 2b2 j  ...  ai n bn j

3. Prove by writing out in full

  x x j  xi  j 1 x j
10 5 5
a. k 1
[(k  1)3  k 3 ]  113  1 b. j 1 i

c.  
3
i 1
2
j 1 i
x xj   
2
j 1
3
x xj
i 1 i
d.  
3
i 1
2
j 1 i
x xj    x  
3
i 1 i
2
j 1
xj 
 x x  x  i 1 xi where x  13  i 1 xi
3 3 3
e. i 1 i

4. Prove using the rules of summation:

     
n i n n n n
a. i 1 j 1
i  i 1
i2 b. i 1 j i
j  i 1
i2

5. Prove all the equality relations in the following:

n n n

 (x
i 1
i  x )2   (x
i 1
i  x ) xi  x i 1
i
2
 nx 2

n n
6. Show that  (x
i 1
i  x )( yi  y )   xi yi  n x y .
i 1

Matrix Algebra Notes 3-6

Anthony Tay
7. Evaluate by first simplifying, then applying the appropriate formulas

 
30 n
a. k 1
k (k  2) b. k 1
k (k  2)(k  2)

8. Let { x1 , x2 ,..., x10 }  {1, 2, 6, 13 ,  , e, 0,  1, 2, 4.2 } . Verify using excel or

otherwise (calculator, manual computation, mental computation, as you
please) that


10
a. i 1
( xi  x )  0

 ( xi  x ) 2   i 1 xi2  10 x 2
10 10
b. i 1

 ( xi  x ) 2   i 1 ( xi  x ) xi
10 10
c. i 1

If in (i) you get an answer a little different from 0, explain why this occurs.

9. Express the following in the form ax1  bx2  cx3  dx4 :

   
4 2 4 4
(i) i 1 j 1
i xj (ii) i 1 j i
i xj

(i.e., you have to tell me what a , b , c , and d are in each case.)

10. Let {x1 , x2 ,..., xn } be an arbitrary set of n real numbers, and let

 i 1 xi .
n
x 1
n
Prove that
 i 1 ( xi  x )( xi  1)  i 1 ( xi  x )( xi  10000) .
n n


Matrix Algebra Notes 3-7

Anthony Tay