1

MATHEMATICS
SELECTED TOPICS FOR THE STUDENTS OF FACULTY OF PHARMACY

Petr Klemera , Michaela Hamerníková

Hradec Králové 2008

2
 3

Introduction
This text book is dedicated especially to students of faculty of pharmacy but it can be useful
also for students or workers in other disciplines, namely in biomedicine or biochemistry.
Selection and the range of content is focused especially on the explanation of the meaning of
mathematical ideas with regard to the practical application.
The first two chapters concern solutions and mixtures and calculation with inexact numbers,
third, fourth and fifth chapter repeat and extend knowledge of secondary school concerning
functions of one real variable and the application of the most important concepts of
mathematical analysis – the derivative and the integral. Sixth chapter briefly deals with the
functions of more variables. Seventh chapter acquaints with the main ideas and the use of the
theory of matrices.
4
 5

Content
1 SOLUTIONS AND MIXTURES ....................................................................................... 7
1.1 Two-part solutions and their mixtures ......................................................................... 7
1.2 Three- and more-part solutions and their mixtures ...................................................... 9
2 CALCULATIONS WITH INEXACT NUMBERS ......................................................... 12
2.1 Addition and subtraction of inexact numbers ............................................................ 13
2.2 Multiplication and division of inexact numbers ........................................................ 14
2.3 Estimate of the precision of function value ............................................................... 15
3 FUNCTIONS OF ONE VARIABLE ............................................................................... 17
3.1 Basic concepts ............................................................................................................ 17
3.2 Linear functions ......................................................................................................... 18
3.3 Power functions ......................................................................................................... 19
3.4 Exponential functions ................................................................................................ 19
3.5 Logarithmic functions ................................................................................................ 20
3.6 Goniometric functions ............................................................................................... 22
3.7 Graphic solution of equations .................................................................................... 23
3.8 Identification of some types of functions .................................................................. 24
4 THE DERIVATIVE ......................................................................................................... 27
4.1 The definition and geometric meaning of the derivative ........................................... 27
4.2 Physical meaning of derivative .................................................................................. 29
4.3 Derivative formulas ................................................................................................... 29
4.4 Course of function and local extremes ...................................................................... 32
5 THE INTEGRAL ............................................................................................................. 37
5.1 Definition and properties of the indefinite integral.................................................... 37
5.2 Integration formulas ................................................................................................... 38
5.3 Definite integral ......................................................................................................... 41
5.4 Properties of the definite integral ............................................................................... 42
5.5 Differential equations ................................................................................................ 45
6 FUNCTIONS OF MORE VARIABLES.......................................................................... 47
6.1 Basic concepts ............................................................................................................ 47
6.2 Functions of two variables and their representation .................................................. 47
6.3 Partial derivative ........................................................................................................ 48
6.3.1 Definition and calculation of the partial derivatives .......................................... 48
6.3.2 The partial derivative and the course of functions ............................................ 50
7 MATRICES ...................................................................................................................... 53
7.1 Matrix operations ....................................................................................................... 53
7.2 Unit and inverse matrices .......................................................................................... 55
7.3 Application of inverse matrix to the solution of systems of linear equations............ 56
6
 7

1 SOLUTIONS AND MIXTURES

Calculations concerning solutions and mixtures of compounds should be the routine matter in
pharmacy with emphasis to the correctness of the results. The danger coming from incorrect
results in pharmacy is obvious because it concerns the health of the patient. To minimize the
risk of the mistake, it is appropriate to use a standard, generally valid method for all tasks of
this type.

1.1 Two-part solutions and their mixtures

The two-part solution is a solution containing solvent and one substance dissolved in it. The
two-part mixture is the mixture of two substances. Let´s presume that substances do not
undergo chemical reaction in all cases.

Symbols used here:

Mass of solution or mixture: M

Mass of substance or component: m
m
Concentration: c = (1)
M
Concentration (per cent) C = 100 c (%)

Remark: Definition (1) is "mass concentration". Other definitions used in practice are, e.g.,
"mass-volume concentration" m/V or "volume concentration" v/V , where V is volume of
the solution, v volume of the substance. The usage of these concentrations for calculations is
similar to the calculations described further, but only for small values of concentrations. The
reason is that simple basic equation V1 + V2 + . . . + Vk = V does not generally hold for
large concentrations.

Derivation of mixing rule (two components)

General situation: a number k of parts are mixed. If the result is solution, any of these
parts can be given substance or given solvent or solutions of them. If the result is mixture, the
parts can be substance A or substance B or their mixtures. Let us distinguish masses and
concentrations of individual parts by subscripts 1 , 2 , ..., k , while the mass and concentration
of the resulting solution or mixture have no subscript. All the data representing the general
situation are shown in the table:

p a r t s result
M1 , m1 , c1 , C1 M2 , m2 , c2 , C2 .............. Mk , mk , ck , Ck M,m,c,C
 8

Two obvious equations describe the situation:

M1 + M2 + . . . + Mk = M (2)
m1 + m2 + . . . + mk = m (3)
From definition (1), it follows m = c M and substitution into equation (2) gives the mixing
rule in the following form:
c1 M1 + c2 M2 + . . . + ck Mk = c M (4)
Finally, we can multiply equation (4) by number 100 and (because C = 100c , Ci = 100ci )
we can write both equations in the most useful form

M1 + M2 + . . . + Mk = M
(5)
C1 M1 + C2 M2 + . . . + Ck Mk = C M

The system of equations (5) makes it possible to calculate any problem for solutions of one
substance in one solvent or for mixtures of two chemically different components. It is evident
that there must be exactly two unknown quantities for unique resolution of this system.
Example 1.
We mixe two solutions:
[1] 150 g with concentration 2 %
[2] 50 g with concentration 4 % .
What is the mass and concentration of resulting solution?
Substitution of given values into equations (5) gives
150 + 50 = M i.e., M = 200 g,
2 ´ 150 + 4 ´ 50 = C ´ 200 i.e. C = 2.5 % .
Example 2
We are to prepare 100 g solution of 5 % concentration. We have solvent and pure
substance, but we want make use of 40 g 2 % solution prepared before. That is, we are to use
these parts:
[1] M1 = 40 g 2%
[2] M2 = ? g 0 % (solvent)
[3] M3 = ? g 100 % (substance)
Substitution into equations (5) gives
40 + M2 + M3 = 100
2 ´ 40 + 0 ´ M2 + 100 ´ M3 = 5 ´ 100 i.e., M3 = 4.2 g
and substitution of this to the first equation
40 + M2 + 4.2 = 100 gives M2 = 55.8 g
 9

Exercise 1
How much solvent is it necessary to add to 30g of 5% solution to prepare 3% solution, and
what is the total mass of the resulting solution?
[20g of solvent, total mass 50g]
Exercise 2
To 200g of 20% solution the second solution of the same substance was added to give rise
500 g of 14% solution. What was the concentration of the second solution?
[10%]
Exercise 3
What amount of the pure substance and how much of solvent is it necessary to add to 40g
of 6% solution to prepare 50g of 10% solution?
[2.6 g of the pure substance, 7.4 g of solvent]
Exercise 4
How much 2% and 10% solution is it necessary to mix to prepare 200g of 3% solution?
[25 g of 10% solution, 175 g of 2% solution]

1.2 Three- and more-part solutions and their mixtures

Derivation of mixing rule (three components)

Derivation will be demonstrated on three-part solutions, generalization for more-part solutions

is obvious. When mixing three components (say - substance A, substance B and solvent), we
must define two types of concentrations, one for each substance:

mA mB
cA = cB =
M M (6)
where mA is mass, cA concentration of substance A ,
mB is mass, cB concentration of substance B ,
M is mass of solution or mixture.

Now three independent equations hold:

M1 + M2 + . . . + Mk = M (7)
m1A + m2A + . . . + mkA = mA (8)
m1B + m2B + . . . + mkB = mB (9)

Similarly as above, from (6) it follows mA = cA M , mB = cB M and substitutions into

equations (8) , (9) give system of three equations making possible to resolve any problem
 10

of mixing three components. It is useful again to use percentage version of concentrations, i.e.
CA = 100 cA , CB = 100 cB . There must be exactly three unknown quantities for unique
result of this system.

M1 + M2 + . . . + Mk = M
C1A M1 + C2A M2 + . . . + Ck AMk = CA M (10)
C1B M1 + C2B M2 + . . . + CkB Mk = CB M

It is easy to see that analogous derivation would give a system of four equations for
four components (e.g. solvent and substances A , B , C), five equations for five components,
etc. The number of unknown quantities must be generally the same as the number of
components.
Example 3
We mix following three parts:
[1] 200 g solution with 20 % concentration of substance A (and no substance B)
[2] 100 g solution with 10 % concentration of substance B (and no substance A)
[3] 700 g solvent
What are concentrations and mass of resulting solution?
Substitution of given values into equations (10):
200 + 100 + 700 = M i.e. M = 1000 g
(for A:) 20 ´ 200 + 0 ´ 100 + 0 ´ 700 = CA ´ 1000 i.e. CA = 4 %
(for B:) 0 ´ 200 + 10 ´ 100 + 0 ´ 700 = CB ´ 1000 i.e. CB = 1 %
Example 4
What amounts of substance A, substance B and solvent are we to add to 100 g of solution
where concentration of A is 5 % and concentration of B is 2 %, if we want 500 g of solution
with concentrations 10 % A , 4 % B ?
Given data in more lucid form:
[1] M1 = ? 100 % A
[2] M2 = ? 100 % B
[3] M3 = ? solvent (0 % A , 0 % B)
[4] 100 g 5%A, 2%B
______________________________________
[result] 500 g 10 % A , 4 % B
Substitution into equations (10):
M1 + M2 + M3 + 100 = 500
(for A:) 100 ´ M1 + 0 ´ M2 + 0 ´ M3 + 5 ´ 100 = 10 ´ 500 i.e. M1 = 45 g
(for B:) 0 ´ M1 + 100 ´ M2 + 0 ´ M3 + 2 ´ 100 = 4 ´ 500 i.e. M2 = 18 g
and substitution to the first equation gives 45 + 18 + M3 + 100 = 500 i.e. M3 = 337 g
 11

Exercise 5
part I is solution containing 4% of substance A and 8% of substance B
part II is the pure substance A
part III is solvent
What amounts of these parts are we to mix to prepare 100g of solution containing 4.5% of
compound A and 2% of compound B?
[25g solution, 3.5g of A, 71.5g of solvent]
Exercise 6
part I is 7.5 g of 20% solution of the substance A
part II is 10% solution of the substance B
part III is solvent
What amounts of parts 2 and 3 are we to add to the part I to prepare solution containing 1% of
the substance A, 2% of the substance B and what is the total mass of the resulting solution?
[30g of solution B, 112.5 g of solvent, in total 150 g].
Exercise 7
part I is solution containing 2% of the substance A, 5% of the substance B
part II is solution containing 10% of substance A, 1% of substance B
part III is the solvent
Calculate amounts of these three parts that are necessary for the preparation of 400g of
solution, which contains 0.5% of substance A, 2.45% of the substance B.
[M1 = 200g, M2 = -20g, M3 = 220g, impossible in practice]
 12

2 CALCULATIONS WITH INEXACT NUMBERS

In practice, numbers expressing values of various quantities do not express, as a rule,

exact values of those quantities. The reason is an unavoidable imprecision of measurement of
any continuous quantity (e.g. length, mass, velocity, voltage .....). It is important to know the
level of this imprecision and to express it not only in making notes of these values, but also in
formulations of numerical results of calculations based on such inexact numbers.

Basic concepts

Normalized form of a number X :

X = a ´10 k (11)
where 1 £ a< 10 and k is a whole number (integer) .

Examples
276.3 = 2.763 ´ 102 0.00075 = 7.5 ´ 10-4 5000000 = 5 ´ 106

Number of decimal places of a number

Examples
75.3 has 1 decimal place
75.300 has 3 decimal places
0.0304 has 4 decimal places

Significant digits of a number are all digits beginning with the first non-zero digit from the
left.

Examples
75.3 has 3 significant digits
75.300 has 5 significant digits
0.0304 has 3 significant digits

Rounding numbers
The last digit in the rounded number is determined so that the difference betweeen the
original number and the rounded number is minimum.
 13

Example
Rounding number p = 3.1415925... to 4 decimal places gives 3.1416
to 3 decimal places gives 3.142
to 2 decimal places gives 3.14

A special problem appears for rounding last digit 5 to the preceding digit. For example, we
can write either 7.65 =& 7.6 or 7.65 =& 7.7 and the difference is 0.05 in both cases. The
most correct possibility would be to choose the way of rounding "up" or "down" at random,
for example by the help of tossing a coin. The aim is to avoid so called systematic error in
case of large computations. The same can be reached by “roud-to-even” method: last digit 5
is rounded so that the new last digit is even, e.g. 7.65 =& 7.6 , 7.75 =& 7.8 .

Notation of inexact numbers

There are two common ways of notation of inexact numbers:

a) Explicit notation of the value of error, e.g.
20.0 ± 0.4 (i.e. absolute error is 0.4)
or, equivalently 20.0 ± 2 % (i.e. relative error is 0. 4 ´ 100 = 2 % )
20. 0
b) Expression of the precision by the proper number of significant digits only, e.g.:
20.0 20.00 20.000

The exact meaning of the numbers expressing errors is not unique in practice. Sometimes
(rarely) the number means maximum possible value of the real error. As a result of a
statistical evaluation, the number usually means the value of standard deviation.
Nevertheless, there is no exact interpretation of the value of error in most cases and then the
notation (b) is used in common. Number, say, 25.3 then means that we know "something"
about the value of the first decimal place but nothing about second and next decimal places.
The further rules for determination of number of significant places are formulated so that the
real error should be less than 5 units of the last significant digit (i.e. less than 0.5 for inexact
number 25.3 , less than 0.005 for number 20.000 etc.), but this level of exactness cannot be
fully guaranted.
2.1 Addition and subtraction of inexact numbers
The theory easily shows that the maximum value of error in the result of addition or
subtraction is the sum of absolute errors of numbers used here. This yields to a simple rule:
The number of significant digits in a result of addition or subtraction is determined by the
minimum number of decimal places in all added or subtracted numbers.
 14

Example 5 0 . 0 1 0 3
+ 0 . 0 2 0
+ 0 . 0 0 5 4 1

= 0 . 0 3 5 7 1 @ 0.036

Example 6
2.768 ´ 10 -9 - 3.1 ´ 10 -10 = 10 -9 (2.768 - 3.1 ´ 10 -1 ) =
= (2.768 - 0.31) ´ 10 -9 = 2.458 ´ 10 -9 @ 2.46 ´ 10 -9
Take notice of the necessity of comparison of decimal places for numbers free of
10n-multiplier, as the comparison of decimal places is confusing for different 10n
multipliers.

2.2 Multiplication and division of inexact numbers

According to theory, the sum of relative errors determines maximum of relative error in the
result of multiplication or division. As far as the relative error of an inexact value is
approximately expressed in the number of its significant digits, the following rule holds:
The number of significant digits in a result of multiplication or division is determined by the
minimum number of significant digits in multiplied or divided numbers.
Example 7
{ 21
.2
484 .25 = 0.621 =& 0{
3 ´0{ .2 ´10 -1
.62 = 6{
4 2 2 2
number of significant digits

Example 8
6.4362 ´ 10 3
-5
= 3.2181 ´ 10 8 @ 3.22 ´ 10 8
2.00 ´ 10

Computation of more complicated expressions

We must apply the above described rules for results of each operation in more complicated
expressions. The best method here is to mark significant digits in each partial result, but to
calculate further with more digits and round the final result only.
Example 9 (significant digits are underlined in partial results)
1.111 ´ 0.60 + 2.222 ´ 0.30 = 0.6666 + 0.6666 = 1.3332 @ 1.33

The same calculation where we round each result immediately can give (and really gives
here) little different but acceptable result:

1.111 ´ 0.60 + 2.222 ´ 0.30 = 0.67 + 0.67 @ 1.34

 15

2.3 Estimate of the precision of function value

The basis for practical estimate of the precision of function value is numerical computing of
value Dy = f (x + Dx) - f (x) for given function f and for concrete values x , Dx The
usage of the method is explained in following examples (A) to (D).

(A) The number of significant digits of log(14.6) is to be determined. The error of

value 14.6 might be several units of its last significant decimal place, i.e. several tenths.
To standardize our method, we choose Dx = 0.1 (i.e. one unit of the last significant
decimal place) and compare both function values: log 14.6 = 1,1643...
log 14.7 = 1.1673...
i.e. Dy = 0.0030...
The standard criterion for the decision what are significant digits is as follows:
Significant decimal places of value y are all those where five units of the decimal place
are greater than the difference Dy .
In our example D y = 0.003 < 0.005 (but 0.003 > 0.0005), hence, the thousandths are
the last significant digit and the result should be written as log 14.6 @ 1.164 .
(Note that subtraction instead of addition of Dx gives the same result:
log 14.6 = 1.1643... log 14.5 = 1.1613... again: Dy = 0.0030...

(B) ln z = 2.08 i.e., z = e2.08 = 8.00446…

e2.09 = 8.08491…
Dy @ 0.08
and because 0.08 > 0.05 , the hundredth is the first non significant decimal place and the
result should be rounded as e2.08 @ 8.0 .

(C) 8.277 / 3.00 = 2.759 = 1.66102... @ 1.661

2.769 = 1.66403...
Dy @ 0.003
Note: One unit of the last significant digit (0.01) was added, i.e. next digit(s) have to be
preserved in the computation – other way the number 2.76 would enlarged only 0.001
as compared with 2.759 !

(D) log 251.18 = 2.3999851... @ 2,39999

log 251.19 = 2.4000023...
Dy = 0.0000173...
Pay attention! Just intuitive estimate of the difference (here probably of order 0.1)
might be completely wrong: the true difference is of order 0.00001.
 16

Exercise 8 Calculate and round the result properly

0.4000 ´ 109 - 2.4 ´ 105 [3.998×108]

27.0 ´ 10-12 + 3.6 ´ 10-13 [2.74×10-11]

1.00 ´ 105 - 1.00 ´ 106 [-9.0×105]

1.0 ´ 10 -11 ´ 2.66 ´ 10 -2 [2.7×10-13]

3.752×103 / (2.00×107) [1.88×10-4]

ln(0.065) [-2.73]

log(2.00×107) [7.301]

9.54 [3.089]

3.200 × ln(2.76 + 3.4) [5.82]

log K= (0.744 - 12.00) [K=5.5×10-12]

300.00 ´ 2-1.340 (number 2 is exact) [118.5]

15.226×10-5 – 12.3×10-4 [-1.08×10-3]

8.3×108 + 22.55×1010 [2.263×1011]

12.662×3.33/5.8243 [7.24]

1.200×15.2×105/(12.0000×14.2×108) [1.07×10-4]

ln(5.3 - 0.244) [1.62]

15.3×log(2.000×18.5) [24.0]

8,444 [2.9059]

logK = 1.826 [K = 67.0]

(18.33-0.526) / ln18.233 [6.132]

log15.66 = K×2.57 [K = 0.465]

30.26 (number 3 is exact) [1.33]

 17

3 FUNCTIONS OF ONE VARIABLE

3.1 Basic concepts

Definition. Function f of one variable is a unique mapping of a set Df of real numbers

(called domain of f ) into a set Vf of real numbers (called region of f ).
This definition means that the function determines exactly one number f ( x ) Î Vf
called value of f for each real number x Î Df .
Example 10.
Expression x 2 +1 defines function f with domain Df = R (all real numbers).
Then, it holds for this function e.g, :
f (0) = 0 2 +1 = 1
f (2) = 2 2 +1 = 5
f (-1) = (-1) 2 + 1 = 2 etc.

Definition. Graph of function f is the set of points [x , f(x)] for all x Î Df .

10
9
8
7
6
5
4
3
2
1
0
-3 -2 -1 0 1 2 3

Fig. 1. A part of the graph of function x2 + 1 . Points from Example 10. are marked by little
squares.

Inverse function

Let us remember the concept of inverse function. Let us sign a given function f and its values
f(x) = z . Then inverse function g is defined so that g(z) = x holds for any x Î Df . This
definition makes sense only if the original function f is one-to-one function, i.e. if
f ( x1 ) = f ( x2 ) implies x1 = x2 . If this is not true, the inverse mapping is not unique one.
In this case, it can be possible to define the inverse function for restricted domain of
function f.
 18

Example 11
Function x2 + 1 plotted in Fig. 1 is not one-to-one function on its domain Df = R . It is easy to
see because, e.g., for x1 = 2 , x2 = -2 it holds x12 + 1 = x22 + 1 = 5 . But if the domain is
restricted to x ³ 0 , this restricted function x2 + 1 will be one-to-one . We can now derive
expression describing the inverse function by the help of above definition:
x2 + 1 = z implies either x = z - 1 or x = - z - 1 . But as far as we have accepted
restriction to x ³ 0 only, the adequate solution is x = z - 1 (see alsoFig. 2.

x
(z)
4

0
0 1 2 3 4 z (x)

Fig. 2. Graph of function x2 + 1 restricted to x ³ 0 (thin line and names of coordinates in

brackets) and of its inverse function z - 1 (thick line and names of coordinates without
brackets). Notice that (by the definition of inverse function) if a point [a , b] is in one graph,
the point [b , a] is in the other graph necessarily.

3.2 Linear functions

One of the simplest functions is function of the form f(x) = k x + q , where k , q
are constants. Each function of this form is called linear function. The meaning of
constants k (called slope) and q (called intercept) explains Fig. 3. It is clear that linear
functions are defined for all real values x .

a
q

0 x

Fig. 3. Graph of a linear function f(x) = k x + q is a straight line. It is easy to see that the
intercept q determines the point where the straight line intersect the axis y : that is because
f(0) = k ´ 0 + q = q . The slope k determines the direction of the line: k = tg(a) .
 19

3.3 Power functions

a
The simplest form of these functions is x , where a is a constant. The domain of these
functions depends on value a , but all these functions are defined for non-negative values x .
Some of them (e.g. x 2 , x 3 or 3 x º x 1 / 3 ) are defined for all real values x , some (e.g. x -2 )
with the exception of zero value and some others (e.g. x = x 1/ 2 ) for non-negative
values x only. Generalization of the basic form (with the same domain) can be reached by
multiplication of it with the constant b, resulting in the function in the form b xa.
Examples of graphs of simple power functions are shown in Fig. 4 .

y
x2

x 1.5

x1
x 0.5

0
0 0,2 0,4 0,6 0,8 1 1,2 1,4 1,6 x
a
Fig. 4 Graphs of some power functions x (for positive values a). All of them have two
common points: [0 , 0] (because 0a = 0) and [1 ,1] (because 1a = 1). For exponents a > 1
the curves increase more rapidly for increasing x , while the opposite is true for 0 < a < 1 .
In case of a = 1 , the function is x1= x , i.e. linear function with unit slope and zero intercept.
The graphs of generalized functions b xa have similar form, but all of them have common
points [0 , 0] and [1 ,b] (because b.1a = b ).

3.4 Exponential functions

In power functions, the variable is powered to a constant exponent. On the contrary, in
exponential function a constant is powered to the variable. The simple form of exponential
function is ax where constant a is base of the function. The domain of these functions is
R (all real numbers). Power functions are not defined for negative or zero base. Some
examples of exponential functions are shown in Fig. 5.
 20

y
4
x
2

1.5 x
2

1 1x

0.5 x

-2 -1,5 -1 -0,5 0 0,5 1 1,5 2

x

Fig. 5. Graphs of some exponential functions ax. All of them have one common point [0 , 1]
(because a0 = 1). Notice that exponential curves increrase for a > 1 but decrease for a < 1 .
In the special case a = 1 , the function is 1x = 1 , i.e. function with constant value y =1
(straight line marked 1x ). Notice also the symmetry of functions for base 2 and base
0.5 = 1/2 . That is true because equality ax = (1/a)-x holds. In our case, 2x = (1/2)-x for any
value x.

Special exponential function is ex with the base equal to Euler number e = 2.7128….. .
Similarly as in case of generalized power functions, it is possible to consider generalized
exponential function b ax, where b is an arbitrary constant. For b > 0 graphs of these
functions qualitatively correspond with the graphs in the Fig.5, only the scale on the vertical
axis is modified so that all the graphs cross the point [0,b]. (However, for b < 0 the curve
“turns” down symmetrically according to horizontal axis.)
The same rules as for powers with integer exponents hold for exponential functions, namely:
ax ´ a u = ax +u (12)
ax
= ax -u
au (13)
x u
(a ) = a ux
(14)
x
(ab) = ab x x
(15)
3.5 Logarithmic functions
Logarithmic functions are inverse functions to power functions. Inverse function to
exponential function ax is called logarithm to the base a and signed as loga . By the
definition of inverse function:
a x = z implies loga z = x (16)

It is clear from Fig. 5 that ax are one-to-one functions for all a > 0 except a = 1 and region
of their values is x ³ 0 for all a ¹ 1 . Consequently, functions loga are defined for any
positive a except a = 1 (see also Fig. 6).
 21

Example 12
a) log 2 8 = 3 because 23 = 8
b) log 0.5 8 = - 3 because ( 0. 5) -3 = (1 / 2 ) -3 = 2 3 = 8

The most frequently used logarithmic function is logarithm to the base 10 , called decimal
logarithm and marked shortly as log (instead of regular log10). The reason is that numerical
computations with the base 10 are the easiest ones.

Example 13
Definition (16) has for a = 10 the following form:

10x = z implies log z = x (17)

Then, e.g., 100 = 1 implies log 1 = 0
101 = 10 implies log 10 = 1
102 = 100 implies log 100 = 2
10-2 = 0.01 implies log 0.01 = -2 etc.
Basic properties of exponential functions (see rules (12) to (15) ) and definition of logarithms
as inverse functions to exponential functions imply some important properties of logarithms,
namely:
log a ( xu ) = log a x + log a u (18)
log a ( x / u ) = log a x - log a u (19)

log a ( x u ) = u log a x (20)

log a a =1 (21)

Equalities (18) to (21) hold for any base a where logarithms are defined.
x
3
log 2(z)

1 log 10 (z)

0
1 2 3 4 5 6 7 8 9 10 z
-1

-2

-3
log 0.5 (z)

Fig. 6. Graphs of some logarithmic functions. All logarithmic functions have one common
point [1 , 0]. These functions increase for bases a > 1 but decrease for bases a < 1 . Notice
the symmetry of graphs of log2 and log0.5 . Compare this property with graphs of functions
2x and 0.5x plotted in Fig. 5.
 22

3.6 Goniometric functions

Let us briefly recapitulate the basic properties of goniometric functions. Fundamental
properties, expressed using the rectangular triangle, are shown in Fig. 7.

z
y

Fig. 7. If x, y are the legs and z is the hypotenuse of the rectangular triangle, for goniometric
functions it holds: sin a = y /z, cos a = x / z, tg a = y / x, cotg a = x / y.

Rectangular triangle is not usable for demonstration of goniometric functions of angles

greater then right and for negative angles. Ideal geometric tool for arbitrary angle is the unit
circle, see Fig. 8 a,b and explanation below.

sin
a tg a
1 b 1
sin a

a b a
cos b b
-1 0 cos a 1 cos -1 0 g 1

sin b tg g = tg b

-1 -1

Fig. 8. Determination of the goniometric functions using the unit circle.

For functions sine and cosine the radius of the unit circle is the hypotenuse of the
demonstrated triangle, that is its length is equal to 1 and therefore the sine and cosine are
 23

directly equal to the length of the corresponding leg, respectively. If we read the values of
sine or cosine as the vertical or horizontal, respectively, ordinate of the crossing point of the
arm of the angel with the unit circle, we get the correct values of these functions including the
sign. We can see at the picture that, approximately, sin a = 0.85; cos a = 0.5;
sin b = -0.5; cos b = -0.85.
Analogically for the function tangent, the adjacent leg of the rectangular triangle has the
length 1, so that the value of tangent is directly equal to the opposite leg. If we read the values
of tangent as the vertical ordinate of the crossing point of the arm of the angle with the
vertical line going via value 1 on the horizontal axis, we get the correct values of tangent
including the sign. For those arms of angles which do not cross the mentioned line it is
necessary to continue the arm (see value of tg b). For the function cotangent we can use the
general equation cotg x = 1/tg x.

3.7 Graphic solution of equations

Exact analytical solving of some equations is difficult or even impossible. However, in case
of equations of one unknown quantity, it is always possible to find approximate solutions by
the help of numerical or graphic methods. A simple graphic method, illustrated in Fig. 9,
is as follows: It is possible to write any equation of one unknown x in the form
f ( x ) = g( x ) (22)
and solution of this equation are all such values x1 , x2 , . . . for which values of functions
f , g are equal, i.e. [xi , f(xi)] = [xi , g(xi)] holds for points of their graphs. In other words,
intersections of graphs of functions f , g determine the solutions of equation (22).

y
y = x -1

y1 y = 2 -x

0
-1 1 x1 2 3
x

-1

Fig. 9. Graphic solution of equation 2 - x = x - 1 . The unique solution is x1 =& 1.4 .

 24

3.8 Identification of some types of functions

In practice we sometimes meet the problem to determine function describing the relationship
between two variables x and y, values of which are obtained experimentally. We are searching
for the function whose graph corresponds well to all experimental points [x,y].
According to the theory of functions, each n-tuple of such points can be inlaid by the
graphs of infinitely many functions, so that task formulated this way has no good meaning.
Therefore, it is necessary to constrict the choice of functions to several chosen or presumed
types. Then the problem is to recognize if the given distribution of the points corresponds to
the chosen function type, and if yes, to determine the concrete parameters of this function, in
other words to identify this function.
The easiest possibility for the identification represent those functions which graphs are
possible to linearize, it means to find such ordinates, in which the graphs of all functions of
this type are straight lines.

Example 14
Generalized power functions b xa have two parameters, constants a, b. Let us presume there
really exists relation y = b xa between some variables y and x. Applying the logarithm on
both sides of this equality we get the equation log y = log b + a.log x. If we formally consider
z = log y, u = log x, we get the linear relation between the new variables z, u: z = log b + a.u.
In other words – in logarithmic ordinates the graph of the function y = b xa is the line with
the slope a and intercept log b (see Fig. 10.

z = logy

u = log x
log b

Fig. 10. In logarithmic ordinates the graph of function y = b xa corresponds to linear

function z = a.u + log b, and its graph is straight line with slope tg a = a and intercept log b.
 25

Example 15
The kinetics of chemical reactions describes the course of concentration changes in time.
These changes can have various character called reaction order. For example, reaction of the
first order is expressed by the relation
c (t) = C0.e-kt (23)
where c is the concentration, t is time, C0 = c(0) is the starting concentration, k is the rate
constant. Let us presume the experimentally obtained values of concentration during the
course of reaction written in the first two rows of the table:

t [s] 0 10 25 45 65
c [mol/l] 57.9 43.5 28.9 16.5 9.5
log c 1.763 1.638 1.461 1.217 0.978

Our aim is to find whether it is the first order reaction and if yes, we are to determine its rate
constant. Applying logarithm on equation (23) implies equality
log c = log C0 - k.log (e).t (24)
which is the linear relation between the new variable z = log c and time t. It means this course
will be linearized in semilogaritmic ordinates (with logarithm only on the vertical axis).

60 1,8
c log c
50 1,6
a b
40
1,4
30
1,2
20
1
10

0 0,8
0 20 40 60 80
0 20 40 60 80
t t

Fig. 11 (see example Example 15).

Graph a illustrates the non linear course of concentration within time, graph b illustrates the
course linearized in semilogarithmic ordinates; this is the proof of the first order reaction.
The curve connecting the points in graph a is the graph of exponential function (23).
Its representation in graph b is the straight line (24) with parameters log C0 = 1.76 ,
k.log(e) = 1.76 / 80 = 0.022. Hence, C0 = 101.76 = 57.5 ; k = 0.022 / log(e) = 0.0507
are the identified values of parameters.
 26

Example 16
Let us presume quadratic function relation without absolute term, i.e. functions of form
y = bx + cx2.
This equation can be adapted to the form
y
= b + c.x
x
which means that this function will be linear in ordinates x , y / x.

Identification of functions is connected with the searching for so called regression functions. These
tasks often arise from typical situations when experimentally obtained values are only approximate,
inexact. The goal then is not to inlay the function relation exactly but only to find such a function of
the considered type which corresponds in a best way to the found distribution of points. In the
examples mentioned above the regression line would have been determined by a statistical
computation called “minimum least square method”.
 27

4 THE DERIVATIVE
4.1 The definition and geometric meaning of the derivative
The concept of derivative is the most important concept of calculus. It expresses an important
local property of functions, namely the local "speed" of changes of f(x) when x is changed
slightly.

Example 17
Let us try to express local "speed" of change of function x2 in the neighbourhood of value
x = 1. The speed of change of this function means the ratio of the change Dy of value y = x2
and the change Dx of value x . For, say, Dx = 0.5 the change of x2 is
Dy = 1. 52 - 12 = 2. 25 - 1 = 1. 25 and the ratio D y / D x = 1. 25 / 0. 5 = 2. 5 . Let us compare
similar calculations also for smaller values of Dx :

Dx 0.5 0.2 0.1 0.01 0.001

(1+ Dx ) 1.5 1.2 1.1 1.01 1.001
(1+ Dx )2 2.25 1.44 1.21 1.0201 1.002001
Dy = (1+ Dx )2 - 12 1.25 0.44 0.21 0.0201 0.002001
Dy / Dx 2.5 2.2 2.1 2.01 2.001

As values Dx come close to zero, values of the ratio D y / D x come close to value 2.
We can easily understand this property from Fig. 12. Take notice of the tangent line to the
graph of function x2 in its point [1 , 1]. The ratio D y / D x is the ratio of sides in right
triangle drawn there. If Dx is smaller and smaller, the triangle is smaller as well and the
direction of its hypotenuse is closer and closer to the direction of the tangent line, marked as
angle a .
The process described here can be expressed in more formal way by the help of mathematical
concept of limit :
Dy
tga = lim , i.e., for y = x 2 ,
D x®0 Dx

(1 + Dx) 2 - (1) 2 12 + 2Dx + Dx 2 - 12

tga = lim = lim = lim (2 + Dx) = 2
Dx ®0 Dx Dx ®0 Dx Dx ®0

and this value 2 of the slope of tangent line is called derivative of function x2 in point x = 1.
 28

y
2,5

1,5
D y

1
a
D x
0,5
0

-0,5 0,5 1 1,5 x

-0,5

-1

-1,5
-2

Fig. 12. Graph of function x2 and its tangent line for x = 1. The limit of ratio D y / D x
for Dx approaching zero is equal to tg a , i.e. to the slope of the tangent line.

Generalization of the explanation given in Example 17 yields to following definition (see

alsoFig. 13): the slope of tangent line to the graph of function f(x) in point [x , f(x)] is
determined by the limit

f ( x + Dx ) - f ( x )
lim = f ¢( x ) (25)
Dx ® 0 Dx

and called derivative of function f(x). The derivative itself is a function mapping uniquely
values x to values f ¢ ( x ) of slopes.

f (x+ D x)

Dy

a
f (x)
Dx

x x+ D x

Fig. 13. Illustration to the definition of derivative f ¢( x) = tga of function f(x) .

 29

Example 18
Similarly as in Example 17, we can derive derivative of function x2 by definition (25):

( x + D x) 2 - x 2 x 2 + 2 xD x + D x 2 - x 2
( x 2 ) ¢ = lim = lim = lim (2 x + D x) = 2 x (26)
D x ®0 Dx D x ®0 Dx D x ®0

Hence, the slope of tangent line to the graph of function x2 is 2x for every real value x .
(Compare this result with Fig. 12 and Example 17.)

4.2 Physical meaning of derivative

Let the variables in Fig. 13 have the following physical meaning: x = t (time) and y = s
(path of a moving physical body). Then the ratio D y / D x , e.i. D s / D t , is mean (average)
velocity during time interval Dt . The limit of this ratio for Dt tending to zero is actual
(instantenous) velocity of the body at time t . Really, the slope of the tangent line shows the
actual speed of path changes. This is expressed in following definition of actual velocity v(t) :

v(t ) = s ¢(t ) (27)

The actual velocity itself can vary with time. The actual speed of its change is expressed as
derivative of v(t) and called actual acceleration or shortly acceleration:

a (t ) = v ¢(t ) (28)

Substitution of (27) into (28) gives equality

a (t ) = ( s ¢(t ))¢ = s ¢¢(t ) (29)

where s ¢¢(t ) is second derivative of function s(t) - nothing else then derivative of the (first)
derivative of s(t) .

Many different quantities can change with time, e.g. concentration, power, pressure etc.
Derivative of such quantity is also called (actual) velocity of changes of concentration, power
etc. In case that some quantity changes with other variable than time (say: atmosphere
pressure changes with elevation, i.e with above sea level), the derivative of such quantity
expressing the actual speed of changes of that quantity with changes of given variable, is
called gradient (e.g. gradient of pressure).
4.3 Derivative formulas
In practice, derivatives are calculated by the help of formulas (as a matter of fact, one of them
was derived in Example 18. It is not necessary to learn these formulas by heart, but to be able
to use them for computing derivatives of various functions.

General formulas
 30

¢
multiplication by constant C [C ´ f ( x)] = C ´ f ¢( x)

addition or subtraction of functions [ f ( x) ± g ( x)]¢ = f ¢( x) ± g ¢( x)

multipliciation of functions [ f ( x) ´ g ( x)]¢ = f ¢( x) ´ g ( x) + f ( x) ´ g ¢( x)

¢
é f ( x) ù f ¢( x) ´ g ( x) - f ( x) ´ g ¢( x)
division of functions ê g ( x) ú =
ë û g 2 ( x)

derivative of compound function [ f ( g ( x) )]¢ = g ¢( x) × f ¢( z )

( z = g(x) must be substituted back after computing derivative f ´(z) )

Special formulas
constant K¢ = 0

power functions ( x a )¢ = a xa -1

exponential functions (e x )¢ = e x (a x )¢ = ln a ´ a x
1 1
logarithmic functions (ln x)¢ = (log a x)¢ =
x x ´ ln a
goniometric functions (sin x)¢ = cos x (cos x)¢ = -sin x

( tg x)¢ see Example 22

Example 19 (in detail for the first time)

¢ ¢ ¢ ¢ ¢
( ) ( ) ( ) ( )
2 x 3 - x + 5 = 2 x 3 - x 1 + (5) = 2 x 3 - 1 × x 0 + 0 = 2 × 3 × x 3-1 - 1 + 0 = 6 x 2 - 1

Example 20
( x ×ln x + 2 x 3 - 5)¢ = ( xln x )¢ + (2 x 3 )¢ - (5)¢ = ( x)¢ × ln x + x × (ln x)¢ + 2 ( x 3 )¢ - 0 =
1
= ln x + x + 2 ´ 3 x 2 = ln x + 1 + 6 x 2
x
Example 21
[sin( x 2
] [ ]
+ 5 x) ¢ = ( x 2 +5 x)¢ ´ (sin z )¢ = ( x 2 )¢ + (5 x)¢ ´ cos z = [2 x + 5] ´ cos( x 2 + 5 x)

Example 22
(sin x) ¢ cos x - sin x(cos x) ¢ cos 2 x + sin 2 x 1
(tg x) ¢ = (sin x / cos x) ¢ = 2
= 2
=
cos x cos x cos 2 x
 31

Exercise 9

a) y = 3x2-5x+2 [ y ¢ = 6 x - 5]
b) y = (2x-3)(1-x) [y ¢ = 5 - 4 x]
3 é ¢ 4 2 3ù
c) y = x5-2y3-4x2+ êë y = 5 x - 6 x - 8 x - x 2 úû
x
d) y = (x2-3)(x2+2x-1) [ y¢ = 4 x 3 + 6 x 2 - 8x - 6 ]
3 - x2 é ¢ 2 ù
e) y = êë y = - 5 x úû
5
2x + 5 é 11 ù
f) y= ê y¢ = -
x -3 ë (x - 3)2 úû
g) y =
x3 é
ê y ¢ =
(
x2 x2 + 3 ù )
2 ú
x2 +1 ëê (x 2 + 1 úû )
x 2 -1 é 3 - x2 ù
h) y = ê y ¢ = ú
x3 ë x4 û

i) (
y = x2 + x +1 ) 2
[y ¢ = 4 x 3
+ 6x 2 + 6x + 2 ]
j) y = (x 2
+ 5x + 7 )
8
[y¢ = 8(2x + 5)(x 2
+ 5x + 7 )]
7

1 é 8x ù
k) y = ê y¢ = - 2 5 ú
(x 2
+1 )
4
êë (
x + 1 úû )
é 2x - 3 ù
l) y = x 2 - 3x + 2 ê y¢ = ú
2
ë 2 x - 3x + 2 û
m) y = xe x [y ¢ = (x + 1) e ] x

n) y = e x - x
2
[y¢ = (1 - 2 x)e ] x- x2

o) y = 2 x - x 2 [2 ln 2 - 2 x, x Î R]
x

p) y = x 3 - 3 x + 1 [y ¢ = 3x - 3 ln 3]
2 x

2 x -1 é 2 x +1 ln 2 ù
q) y = ê y¢ = x 2 ú
2x +1 êë (
2 + 1 úû )
é 2x + 1 ù
r) (
y = ln x 2 + x - 2 ) êë y ¢ = x 2 + x - 2 úû

é 3x 2 ù
s) y = log x - 1 ( 3
) ¢
êy = 3 ú
ë (
x - 1 ln 10 û )
 32

4.4 Course of function and local extremes

As mentioned above, the value f ´(x) of derivative is equal to the slope of tangent line to the
graph of function f at point [x , f (x)]. As tangent line is close to the graph in a small
neighbourhood of that point, the value of derivative expresses the basic „local property“ of the
function near the point. The following Fig. 14 A illustrates three substantially different local
situations.

A f(x)

a b c x

f’(x)

B
c

a b x

Fig. 14. Graph of function f and its derivative.

A) Graph of the function f with tangent lines for x = a, x = b and x = c drawn.
B) Graph of the derivative f ´ of the same function expresses the slopes of tangent lines
to the graph of function f for all corresponding values x.

The function is increasing near point [a , f (a)] and the tangent line is increasing, too. Hence,
derivative f ´(a) is positive. In point [c , f (c)], both the function and the tangent line decrease
and derivative f ´(c) is negative. A special case is the point [b, f (b)]. Here the function has
its local maximum. Apparently, the tangent line is parallel to the x axis and its slope is equal
zero. Hence, f ´(b) = 0 holds. Taking into accout an estimate where the slopes are greater or
smaller, the graph of the derivative can be estimated as shown in Fig. 14 B.

The following Fig. 15 illustrates all possibilities of zero slope of tangent lines. The function
g(x) has local minimum in point [a , g(a)] (or, shortly, in point a ), function h(x) has
„horizontal“ tangent line (i.e., zero derivative) in so called saddle point [a , h(a)]
 33

A
f(x)

a x

B
g(x)

b x

C D
h(x) k(x)

c x d x
Fig. 15 Four possible types of functions with the horizontal tangent line.
A) local minimum B) local maximum C), D) saddle points

As local maximum, local minimum or saddle points represent the only situations with zero
slope of tangent line, the solution of equation f ´(x) = 0 gives all possible values of x
where local maxima or minima can (but need not) be .
 34

According to Fig. 16, it is obvious that in A

f(x)
case of local minimum the derivative in the
left neighborhood of the point b is negative
and in the right neighborhood it is positive.
The slope of the tangent line of the derivative
of this function in this point can be
increasing as it is shown in Fig. 16. But in
extreme case the graph of the derivative a x
could have for x = a also saddle (dotted
B
line), and therefore horizontal tangent line. f´(x)

Analogically in the local maximum the

derivative in the left neighborhood is positive
and in the right one it is negative, and the x
a
tangent line is therefore either decreasing or
horizontal for the saddle-type course of the
derivative.

Fig. 16. A) local minimum of function)

B) possible graphs of derivative f
´´(x)
As a result of this knowledge we can formulate the preposition:
a) If for any function f (x) in some point x = a equality f ´(a) = 0 holds, it means that this
function has in this point either local minimum or local maximum or one of the two
types of saddle points.
b) Distinguishing of these cases is possible only on condition that the second derivative f ´´
is not equal to zero in this point. For f ´´(a) > 0 there is local minimum in the point a
(the tangent line to the graph of the first derivative increases), for f´´(a) < 0 it is local
maximum there. If f´´(a) = 0, it means that it can be any of the four possibilities.

Example 23
Let us determine the local extremes of the function x2 – 3x + 3.
The first step is to find the solution of the equation (x2 – 3x + 3)´ = 0, thus 2x - 3 = 0; x = 1.5.
The second derivative (x2 – 3x + 3)´´ = (2x – 3)´ = 2 is constant function, which has in any
point (including point x = 1.5) the positive value 2. So the function has in the point 1.5 local
minimum and the function value in this point is 1.52 – 3×1.5 + 3 = 0.75.
 35

Example 24
In pharmacokinetics the main interest is paid to the change of the concentration of the active
substance in various parts of organism and to the dependency of these changes on the dosage
and the way of application of the medicine. The easiest example is intravenous injection and
than the monitoring of the concentration of the substance in the blood and in the target organ.
The basic concept is depicted in Fig. 17. In the left side the central compartment (blood bed)
is depicted, in the right side the peripheral compartment (target organ) is drawn; quantities c1
and c2 are time-dependent concentrations of the active substance.

C1 C2

Fig. 17. Two-compartment system with the emanation from the central compartment.
Illustration to the Example 24.

This situation can be mathematically described using a system of differential equations and its
solution for the concentration in the target organ can be derived in form:
c2 (t) = A (a t – e-b t ), where A, a, b are positive constants; a < b.
It is possible to find the values of these constants experimentally. The course of this function
is shown in Fig. 18. .
.

c2(t)

0 t
Fig. 18. Time dependence of concentration of medicine in the target organ
in two-compartment system.
 36

Let us look for the time of the maximum concentration of the active substance in the target
organ. So let us find the local extreme of function c2 (t) by the help of its derivative
(derivative of compound function).
c2 (t ) = [ A × (e - at
)]
- e - bt ¢ = A × b × e - bt - a × e - at = 0
( )
As A ≠ 0, the expression in the brackets must be equal to zero:
(b × e - bt
)
- a × e - at = 0 , i.e. b × e - bt = a × e - at
Applying of the natural logarithm (remember that ln e = 1) implies the solution
ln b - ln a
t1 =
b-a
As the solution is the single one, it is obvious from Fig. 18. that t1 is the maximum, so it is
not necessary to prove it using the second derivative.
Notice that time when the concentration in the target organ is maximum is dependent only on
the parameters a, b (connected with the rate of the transfer between compartments and the
rate of emanation from the system), not on parameter A connected with the dose.

Exercise 10
Find all local extremes of the functions complete with the type of extremes and the function
values in the extremes.

a) 1 x 3 - 7 x 2 + 10 x - 2 [minimum (for x = 5, value 2.1 6 ); maximum (for x = 2, value 6. 6 )]

3 2

é æ ln 5 öù
b) y = e 2 x - 10 x êmin ç 2 , 5 - 5 ln 5 ÷ ú
ë è øû

é æ 5 400 öù
3 2
c) y = x - 2 x - 15 x êmax ç - 3 , 27 ÷ú
ê è øú
êëmin (3, - 36 ) úû

é æ 3 ö ù
êmin ç 3 , 2 3 ÷ ú
ê è ø ú
x3 ê æ 3 öú
d) y = 2 ; x ¹ ±1 êmax ç - 3 , - 3 ÷ú
x -1 ê è 2 øú
ê saddle (0, - 1) ú
ê ú
ë û
 37

5 THE INTEGRAL
5.1 Definition and properties of the indefinite integral

The operation inverse to the derivative is integration.

x2 derivative
x2 +1 2x
2
x -5
... integration
...

Fig. 19. Derivative maps any function of the form x 2 + C , where C is any constant (any
real number), to function 2x . The inverse mapping is called integration and its result is
called integral. Hence, the integral of function 2x is any function of the form x 2 + C
(i.e. the integral is the set of all functions of this form).

The situation shown in Fig. 19 can be written down like this:

( x 2 + C )¢ = 2 x implies ò 2 x dx = x2 + C (30)
The expression ò 2 x dx reads " integral of 2x dx " and represents the set of all functions
of the form x 2 + C for any real constant C (see also Fig. 20). Each of these functions is
called primitive (or antiderivative) of function 2x . Hence, the integral of function 2x
means the set of all primitives of function 2x . Symbol dx used as a multiplier of the
integrated function explicitly expresses the name of the variable; the reason is its use in
substitution method of integration (see later).
8

4
x2 + 3.5

2 x2 + 2.5
1
x2
0
-2 -1 1 2
-1
2
-2 x - 1.5

Fig. 20. Graphs of several functions of the form x 2 + C . The value C determines only
vertical shift of the graph.
 38

5.2 Integration formulas

Similarly as in case of derivatives, integrals of some functions can be calculated by the help of
formulas. In fact, these formulas can be derived from derivative formulas. It is based on the
following implication (compare with expression (30) ):

[ f ( x) + C ] ¢ = f ¢( x) implies ò f ¢( x) dx = f ( x) + C (31)

Let us remember again: symbol dx used as multiplier of the integrated function explicitly
expresses the name of the variable only; the reason is its use in substitution method of
integration (see later).

General formulas

multiplication by constant K ò K f ( x) dx = K ò f ( x) dx
´ ´

addition or subtraction of functions ò ( f ( x) ± g ( x)) dx = ò f ( x) dx ± ò g ( x) dx

(there are no general rules for integral of multiplication and/or division of functions)

Special formulas
constant ò K dx = Kx + C

a x a +1
power function (except x-1 º 1/x ) ò x dx = +C (for any a ¹ -1)
a +1
1
function 1/x ò x dx = ln x + C

x x x ax
exponential functions ò e dx = e + C ò a dx = ln a
+C

logarithmic function ò ln x dx = x (ln x- 1) + C

goniometric functions ò cos x dx = sin x + C

ò sin x dx = - cos x + C

1
ò cos 2
dx = tg x + C
x
Example 25.

ò (5 x
3
- 2x ) dx = ò 5x
3
dx - 2
ò x dx = 5ò x 3 dx - 2 ò 1x dx = 5 x 4 - 2ln x + C
4
Example 26.
x 2xx
ò (1 + 2 ) dx = ò 1 dx + ò 2 dx = x + ln 2 + C
 39

Contrary to the general derivative formulas that enable to determine the derivative of any
function composed of definite number of functions with known derivatives, the situation
concerning integration is much more complicated. In general, there does not exist universal
instructions how to find the integral of any function composed from functions with the known
integral. If the direct integration formulas are not enough, two basic methods can be used:
integration using substitution and integration per partes.

Substitution method is formally derived from the rule for derivative of compound function.
Let us follow the example:

Example 27
We are to determine ò cos(3 x + 7 ) dx . Suitable substitution is z = 3 x + 7 . We will replace the
function in the brackets and the differential dx. First, we need to find the derivative of the
dz ¢
substitution equation: = (3 x + 7 ) = 3
dx
and from that we formally calculate dx = dz 3 . Now we replace substitutes and then integrate
according to the formulas:
dz 1 1 1
ò cos(3x + 7 ) dx = ò cos (z ) 3
= ò cos( z ) dz = sin ( z ) + C = sin (3 x + 7 ) + C
3 3 3
(at the end we gave back the original variable).

Substitution method in general:

ò f [g (x )]dx = ò f (z ) dx , where first we substitute dx = dz g ¢(x )
and after integration we replace back g ( x ) = z .

Problems can arise when in the expression for dx still remains variable x. For integration of
the new integral, it is essential whether only the new variable occurs there. In opposite case it
is necessary to replace the old variable by the inverse function x = g -1 ( z ) ; but it often leads
to more complicated integral that at the beginning, so that it is not possible to calculate it this
way.
Example 28
We are to calculate ò sin (x 2 )dx . Let us try substitution x2 = z, i.e. dx = dz 2 x .
dz
Because ò sin (x 2 )dx = ò sin (z ) 2 x , we have to replace x = z .
dz 1 sin ( z )
Resulting integral ò sin ( z ) = ò dz is more complicated than the original one, thus
2 z 2 z
this substitution is not suitable method.
 40

Exercise 11.

Calculate:
dx é 2 ù
a) òx ê- + Cú
x ë x û
é1 5 2 3 ù
ò (x )
2 2
b) + 1 dx êë 5 x + 3 x + x + C úû

x2 + x +1 é1 2 ù
c) ò dx êë 2 x + x + ln x + C úû
x

x +1 é2 ù
d) ò êë 3 x x + 2 x + C úû
x dx

e) ò 2(e
x
+ x dx ) [2e x
+ x2 + C ]
æ 1 ö é4 1 2x ù
f) ò ç 2 x + 3 sin x + cos x - 2 x ÷dx ê x x - 3 cos x + sin x - + Cú
è 3 ø ë3 3 ln 2 û

Calculate using substitution method:

3 é1 4 ù
g) ò (3x + 1) dx êë12 (3 x + 1) + C úû

é1 2 ù
ò x(x ) ( )
2 5 6
h) + 1 dx êë12 x + 1 + C úû

x é 1 ù
i) ò (x dx ê- + Cú
2
)
+1
2 2
ë 2 x +1 ( û )
é1 2 ù
j) òx x 2 + 1dx ( 2
)
êë 3 x + 1 x + 1 + C úû

3x + 1 é2 ù
k) ò 3x - 1
dx êë 9 (3 x + 5) 3 x - 1 + C úû

-2 x é 1 -2 x ù
l) òe dx êë- 2 e + C úû

x +3 é 2 x +3 ù
m) ò 2 dx ê + Cú
ë ln 2 û
é1 ù
n) ò (cos 2 x - 3sin 3x )dx êë 2 sin 2 x + cos 3 x + C úû
 41

5.3 Definite integral

Integral defined above is also called indefinite integral. The concept of "definite
integral" is connected with the problem of computing areas limited by graphs of functions.

f(x)

x
a b
Fig. 21. The area of the hatched part of figure is signed A. It is limited by the graph of
function f(x), by axis x and by vertical lines in the end points of interval [a , b].

The problem of computing areas like area A in Fig. 21 is solved in the following basic
theorem, called the fundamental theorem of calculus. First, let us use the following common
notation for the area A :
b
A = ò f ( x)dx (32)
a

Read that symbol as "the definite integral of f from a to b" (or, "the definite integral of f
over interval [a , b]").
Theorem
Let f(x) be a function continuous on closed interval [a , b] and let F ¢( x) = f ( x) holds
on [a , b] . Then equality
b

ò f ( x)dx = F (b) - F (a ) (33)

a

holds.

This theorem has a very deep meaning. It says that exact calculation of an area limited by
graph of a function f reduces to subtracting of two values of its primitive function F. In
other words, computation of areas can be reduced to computation of indefinite integral.
 42

Example 29
Let us compute the area determined by the graph of function x 2 over interval [1 , 2]
(see Fig. 22.) We must find a primitive function F(x) to function x 2 :
x 2
dx = x 3 + C is the set of all these primitive functions. Let us take, say, the simplest
ò 3
one, i.e. F ( x) = x 3 / 3 . In accordance with (33), the area is
2 3
13 8 1 7
òx
2
dx = F (2) - F (1) = 2 - = - = @ 2.33
1
3 3 3 3 3

y
5 2
x
4

0
-0,5 0 0,5 1 1,5 2 x

Fig. 22. Illustration to Example 29. The hatched area is to be calculated.

5.4 Properties of the definite integral

Let us presume that f(x) is continous function at interval [a , b].
b c b
a) If a ≤ c ≤ b, then ò f (x ) dx
a
= ò f (x ) dx + ò f (x ) dx
a c

b) If everywhere in the interval [a , b] f(x) < 0 holds, then the definite integral expresses
the area as negative number.
a
c) ò f (x ) dx = 0
a
b a
d) ò f (x ) dx
a
= - ò f ( x ) dx
b

Connecting properties mentioned above we can conclude that if the graph crosses the
horizontal axe, we must calculate separately the definite integral of each segment given by the
outer points of the interval and all these crossing points.

In pharmacology, the problem of quantitative expression of the total effect of the medicine
often arises. In general, the magnitude of the effect depends on concentration of medicine in
the area of its action and the concentration varies within time. On condition that the effect is
 43

directly proportional to the concentration, the total effect can be expressed as the following
definite integral:
b
U C (a; b ) = ò c(t ) dt
a

On condition that the effect is directly proportional to logarithm of concentration,

corresponding expression is
b
U C (a; b ) = ò ln [c(t )] dt
a

From the practical point of view, it is necessary that the concentration lies in so called
therapeutic range (too low concentrations usually are not efficient and too high concentrations
are harmful). Let us presume the curve of concentration occurs in this range.
In Example 24 we considered two-compartment system, where concentration of medicine in
the target organ is described by the function c2 (t) = A (e-a t – e-b t ). Let us presume the effect
is proportional to concentration. Then
t2

U C (t1 , t 2 ) = ò A (e )
- at
- e -bt dt , where t1 and t2 determine time range for calculation of the
t1

effect. The situation is depicted in Fig. 23

According to general integration formulas and using substitution1 for integration of e - at and
e - bt , respectively, we get the equation for the expression of the effect of the medicine in the
range [t1, t2 ]:
éæ 1 ö - at1 æ 1 ö - at1 -bt ù
U C (t1 , t 2 ) = A.êç -bt
÷ e -e 2 -ç ( ÷ e -e 2 ú ) ( )
ëè - a ø è-bø û
0 -a t
For t1 = 0 and t2 → ∞ this expression simplifies, as e = 1 and lim e = 0 , hence
t ®¥

é1 1ù
U total = A × ê - ú
ëa bû

c2(t)

0 t1 t2 t
Fig. 23 Effect of a medicine (see Example 24). As geometrical meaning of definite integral
is the area under the curve in the given range, the effect of the medicine corresponds
to the hatched area.

dz - at 1 -z 1
1
Let us substitute e-at = z. Then = -a and òe dt = - ò e dz = - e - at + C . Similarly for b.
dt a a
 44

Exercise 12

2
a) ò (2 x + 1) dx [= 4]
1

b) Calculate hatched area shown in the following graph. [ @ 2.885]

6
2x
5

0
0 0,5 1 1,5 2 2,5
x

3
c) Calculate area under graph of function x above interval [0 ; 1]. [= 0.75]
 45

5.5 Differential equations

Let us remember the concept of equation:
4x + 1 = 2x - 5

is an example of (common) equation for unknown variable x . To solve this equation means
to find all values x , fulfilling the equality described by the equation. (In our case, there is
only one solution x = -3.) Another example of such equation can be quadratic equation like
this:
2 x 2 = 30 - 4 x
with exactly two solutions x1 = 3 , x 2 = - 5 .
The common form of equation for one unknown variable x is
f ( x) = g ( x)

where f , g are known (given) functions - see also (22) on page 23.

A different case is the following equation, where f(x) is unknown:

x 2 + 2 f ( x) = 1 - f ( x)

It is impossible to find value x for which this equality holds, because function f(x) itself is
not known here. This is an example of (common) equation for unknown function f(x) , which
has in this case one solution
1- x2
f ( x) = .
3
Differential equation is also an equation for unknown function f(x) , but derivative of that
function must be present in the equation. For example,
1 + f ¢( x) = x 2 (34)

is differential equation for unknown function f(x) .

It is easy to solve equation (34):
3
f ¢( x) = x 2 - 1 implies f ( x) = ò ( x 2 - 1) dx = x - x + C
3
x3
Hence, all functions of the form -x + C (for any real constant C )
3
are solutions of equation (34). This is a general property of differential equations: they have
either infinitely many or no solutions.

Example 30.
Let us have a physical body moving with constant velocity v(t) = v . What is the function
s(t) describing the dependence of path s of the body on time t ?
 46

By definition (27) ,
v = s ¢(t ) (35)

holds in our case. This simple differential equation has solution as follows:
s (t ) = ò s ¢(t ) dt = ò v dt = v ò 1dt = v t + C

Again, our differential equation have infinitely many solutions determined by arbitrary values
of real constant C . What is the physical meaning of this constant? Let us put t = 0 :
s (0) = v ´ 0 + C = C

i.e. the value C is the path of our body in time t = 0 , marked usually as s0 . Hence the
solution of our problem is the well known formula for the path of motion with constant
velocity:
s (t ) = vt + s 0

The solution of equation (34) was possible by the help of integration. This is not possible in
case that not only derivative f ¢(x) , but also unknown function f(x) itself occurs in the
equation, e.g.:
f ¢( x) = 2 f ( x) (36)

Solving of such differential equations is usually much more complicated (and sometimes
unknown or even impossible). That is not case of equation (36) - it is known that this type of
equations (linear differential equations with constant coefficients) have solutions in form of
sum of exponential functions, in our case single function C e k x . The solution is then reduced
to substitution of this (not fully determined) function into the equation:
(C e k x )¢ = 2C e k x , i.e. k C e k x = 2C e k x

The last equality holds for any real constant C , but only if k = 2 . That means the solutions
of equation (36) are all functions f ( x) = C e 2 x for any real constant C.

Let us summarize some basic properties of differential equations. If a differential equation has
a solution, then it has infinitely many solutions determined by different values of some
constant or constants. For determination of one concrete solution, it is necessary to know not
only the differential equation itself, but also another condition concerning the solution.
(In Example 30 it was the condition s (0) = s 0 .) Solution of differential equations where not
only the derivative of unknown function but also the function itself occurs is generally
impossible by the help of direct integration only.
Differential equations are the basic mathematical means for the description of continuous
processes in physics, chemistry, biology and others fields of science.
 47

6 FUNCTIONS OF MORE VARIABLES

This chapter deals with the application of the most important instruments of mathematical
analysis (derivatives and integrals) on functions, which have more than one variable,
especially on functions of two variables.

6.1 Basic concepts

Any ordered n-tuple [a1 , a2 , ... , an] of real numbers ai is called n-dimensional vector.
Let Df be a set of n-dimensional vectors. Function f of n (real) variables determines for
every n-tuple (vector) [a1 , a2 , ... , an] Î Df exactly one real value, called the function value
and symbolized as f (a1 , a2 , ... , an).
For n=1 this definition corresponds to the definition of the function of one variable
(see chapter 3).

6.2 Functions of two variables and their representation

According to definition above, any function of two variables matches each ordered pair of
real numbers belonging to its domain with the real number called the function value. The
principle is depicted in Fig. 24

? [ a,b,c ]

c = f(a,b) y

b ¤ [ a,b ]

a x

Fig. 24. Graphical representation of one value of a function of two variables. The situation
shown in the picture corresponds to function f (x,y) which in the point given by ordinates
x = a , y = b has value f (a,b) = c. This fact is expressed by the point [a, b, c] in
three-dimensional space.

Example 31
The electric power relies on the voltage and the resistance of the appliance (consumer)
according to the relation P = R.I2. In case the both the current and the resistance can change,
this relation represents the function of two variables P = f (I , R ) . From the mathematical
 48

point of view, its domain can be all pairs of the real numbers, but from the physical point of
view the resistance must be non-negative. Part of the graph of this function is in the Fig. 25

0
I

Fig. 25 Part of the graph of the function of two variables defined with the relation P = R.I2

Very useful way of understanding the graphical representation of the function of two variables
is to imagine it as the surface of the landscape, where altitude (elevation above sea-level)
corresponds to the function value and the directions of ordinate axes correspond (say) the
east and the north direction.

6.3 Partial derivative

The derivative of a function of one variable determines the rate of the change of the function
value with the changing value of the independent variable. From geometrical point of view, it
expresses the slope of the graph of the function. For functions of two or more variables, the
situation is much more complicated because the rate and even the sign of the change can
vary according to the fact which variable is changing or how many variables are changing
together.

6.3.1 Definition and calculation of the partial derivatives

As mentioned in chapter 4, for the derivative of function of one variable it is required that the
graph of the function was continuous and without “the sharp breaks”, because only in that
case the tangent line exists there. Dealing with partial derivatives requires functions of two
variables having the surface of the graph continuous and round, without any sharp breaks. It is
possible to determine uniquely the tangent plane in every point of such surfaces. The tangent
plane has analogical role as the tangent line for functions of one variable: tangent plane
 49

provides the best description of the shape of the original surface in the near neighbourhood of
the point of contact (see Fig. 26).

Definition
¶f ( x, y )
The partial derivative of function f(x,y) by variable x is denoted as and defined
¶x
¶f (x, y ) f (x + Dx , y ) - f ( x, y )
as limit = lim (37)
¶x Dx ®0 Dx
The sense of the definition is in the fact that the partial derivative considers only the change
of variable x , while value y is considered as constant. Thus, for practical calculations, only
the knowledge of the calculation of ordinary derivatives is needed.
Similarly, partial derivative by variable y is defined as limit
¶f ( x, y ) f ( x , y + Dy ) - f ( x, y )
= lim (38)
¶y Dy ®0 Dy
where value x is considered as constant.

Example 32
a) Let us determine the partial derivative by x of the function f ( x, y ) = 2 y 2 + x 3
(i.e., let us consider y as a constant)
¶f ¶ ¶ 3
=
¶x ¶x
( )
2y2 +
¶x
( )
x = 0 + 3x 2 = 3x 2

b) Let us calculate the partial derivative of the function by y ( x is now the constant)
¶f ¶ ¶ 3
=
¶y ¶y
( )
2y2 +
¶y
( )
x = 4y + 0 = 4y

Example 33
Let us calculate both partial derivatives of function of two variables f (I , R ) = R.I 2 from
Example 31.
( )
¶ RI 2
=R
¶ 2
( )
I = 2 I .R
¶I ¶I
¶RI 2 ¶
= I2 (R ) = I 2 .1 = I 2
¶R ¶R
Exercise 13: Calculate all partial derivatives of functions
é ¶f ¶f ù
a) f ( x, y ) = x 2 - 2 xy + y 2 ê ¶x = 2 x - 2 y ; = 2 y - 2 xú
ë ¶y û
x é ¶f 1 ¶f x ù
b) f ( x, y ) = ê = ; =- 2ú
y ë ¶x y ¶y y û
x + 2y é ¶f 1 ¶f 2 ¶f 2( x + 2 y ) ù
c) f ( x, y , z ) = ê ¶x = 2 ; ¶y = 2 ; =- ú
z2 ë z z ¶z z3 û
 50

6.3.2 The partial derivative and the course of functions

Value g´(x) of ordinary derivative in point for x = a expresses slope of tangent line to the
¶f ( x, y )
graph of the function corresponding to x = a. Similarly, value for (x , y) = (a , b)
¶x
expresses the slope of tanget line to corresponding point of three-dimensional graph of the
function. This tangent line is located in tangent plane in the direction of x axis (as y value
¶f ( x, y )
was considered to be constant). Similarly, value at the same point expresses the
¶y
slope of tangent line in the direction of y axis (see Fig. 26, Fig. 27.) .

z = f(x,y)
Ty

R
y Tx

a
x

Fig. 26 Tangent lines to the graph of function f (x,y) in point (x , y) = (a , b).

z = f(x,y)
Ty D z2
Dx

R D z1
y Dy Tx

x
Fig. 27. (compare Fig. 26) Line segment Dx is parallel to x-axis, line segment Dy is parallel
to y-axis, line segments Dz are parallel to z-axis. Partial derivatives express the slopes of
¶f ( x, y ) Δ z1 ¶f ( x, y ) Δ z 2
tangent lines Tx , Ty, so that = , = for (x , y) = (a , b).
¶x Δx ¶y Δy
 51

The values and even the signs of both partial derivatives may be different as they express
local behavior of the function in different directions, as illustrated in Fig. 26, where the
derivative by y is positive (line Ty increases with increasing y) but derivative by x is
negative (line Tx decreases with increasing x).

In the case of functions of one variable, tangent lines parallel to x axis correspond to local
extremes or saddle points of the graphs. It is easy to see that tangent planes play
corrersponding role for functions of two variables. Only in case of points of the graph where
tangent plane is parallel to both x and y axes local extremes or saddle poins occur. It means
that both partial derivatives mus be equal to zero in that case (see Fig. 28).

z
z

y a
x

x y

Fig. 28 Local maximum (left) and saddle point (right). The tangent plane, determined by the
two tangent lines is parallel to the plane determined by axes x , y in both cases.

Example 34.
Let us find point(s), where function f (x , y) = –x3 + 6 x y – y3 + 60 has local extreme(s) or
saddle point(s).
¶f ¶f
= - 3x 2 + 6y , = 6 x - 3y 2 so that –3x2 + 6y = 0 and 6x – 3y2 = 0 must hold.
¶x ¶y
System of the two equations has two solutions : x1 = y1 = 0 and x2 = y2 = 2. Hence, tangent
plane is parallel to both axes in points [0 ; 0] and [2 ; 2]. There and only there must be either
local extreme(s) or saddle point(s). (The true situation is shown in Fig. 29).
 52

Local maximum
Saddle point

60

y
3
2
1
0
0
-1
0
1 -1
2
x 3

Fig. 29 Graph of function –x3 + 6 x y – y3 + 60.

Exercise 14
Find point(s), where function g ( x, y ) = x 2 + y 2 - 2 x - 3 y + 14 has local extreme(s) or
saddle point(s).
[ Only one solution: x = 1, y = 1.5 (remark: g (1 ; 1.5) = 10.75 is local minimum).]
 53

7 MATRICES

The concept of matrix is one of the basic concepts of "discrete" or "finite" mathematics, i.e.
mathematics concerning manipulation with finite amounts of individual numbers. Discrete
mathematics does not involve such concepts like continuous functions, derivative, integral
etc., where continual changes are the substantial principle. Matrices are one of the main
meanz for data processing by the help of computers, but the knowledge of their definition and
elementary properties is very important for everybody who wants to formulate problems
concerning large amount of data and /or to use the results of processing them in computers.

Matrix is any rectangular array of numbers enclosed within brackets. The number of rows
and columns of the array determines the size (dimension) of the matrix.

é 2 -3 0 3 ù é1ù é 0 2 - 1ù
ê 1 1.2 0 5 ú [5 2 - 4 1] ê0 ú ê0.5 3 0 ú
ê ú ê ú ê ú
êë- 1 7 8 0.5úû êë3úû êë- 2 1 1 úû

size = rows ´ columns size 3 ´ 4 size 1 ´ 4

size 3 ´ 1 size 3 ´ 3
(row matrix)
(column (square matrix)
matrix)
Individual numbers in the array are called elements of the matrix. The number of row and the
number of columns determin the position of an element in the matrix. These numbers are
written as subscripts in formal description (see matrix A of size k ´ n shown below).

é a1,1 a1, 2 . . . a1,n ù

êa a 2, 2 . . . a 2,n úú
ê 2,1
A= ê . . . . . . ú
ê ú
ê . . . . . . ú
êa k ,1 ak ,2 . . . a k ,n úû
ë

7.1 Matrix operations

Two matrices are equal if they have the same size and if all corresponding elements are
equal.
Example 35
éa bù éu vù a=u b=v
êc d ú = êê w x úú
ú means c=w d=x
ê
êë e f úû êë y z úû e=y f=z
 54

The sum of two matrices of equal size is again a matrix of the same size with elements equal
to sums of corresponding elements.

Example 36
é 5 ù é 2ù é7 ù
ê 2 ú + ê1 ú = ê 3 ú
ê ú ê ú ê ú
êë- 1úû êë4úû êë3úû

Product of a number and a matrix is defined for any number and any matrix. The result is
matrix of the same size and its elements are corresponding elements of the first matrix
multiplied by that number.

Example 37
é2 0 -1ù é6 0 -3ù
3´ ê ú = ê ú
ë 1 -2 3 û ë 3 -6 9 û

Scalar (dot) product is defined for multiplying row matrix 1 ´ n by column matrix n ´ 1
as the sum of products of their elements, as next example shows.

Example 38
é4ù
[2 5 3] × êê-1úú = 2 ´ 4 + 5 ´ ( -1) + 3 ´ 6 = 21
êë 6 úû

Matrix multiplication is a very important operation based on the concept of scalar

product. The product of matrices A , B is a matrix C , where each its element ci , j
is the scalar product of i-th row of matrix A and j-th column of matrix B.

Example 39
é4ù
é2 5 3ù ê ú é21ù
ê1 0 2ú ê-1ú = ê16ú
ë û ê ú ë û
ë6û

Here, the dot product of the first row of the left matrix and the first column of the right matrix
gives 21 - the element in the first row and first column of resulting matrix. Similarly, the dot
product of the second row of the left matrix and the first column of the right matrix gives 16 -
the element in the second row and first column of the resulting matrix.
 55

Example 40
é 3 1ù é12 13 1 14ù
ê2 0ú é2 3 0 5 ù = ê 4 6 0 10ú
ê ú ê6 4 1 -1ú ê ú
êë1 4úû ë û êë26 19 4 1 úû
size 3 ´ 2 size 2 ´ 4 size 3 ´ 4

It is easy to see that the product of matrices AB is defined only if the scalar products are
defined, i.e. if the number of columns in matrix A is equal to the number of rows in matrix
B . (Notice, that product BA in Examples 39, 40 is not defined for this reason.)
In consequence of this, both AB and BA are defined only for “square” matrices of the same
size n ´ n . But even in this case the resulting products are usually quite different, as you can
see in Example 41

. This non-commutativity of matrix multiplication is very important and unusual, because

multiplication of numbers is commutative: xy = yx for any numbers x , y .

Example 41
é1 -1ù é3 2ù é2 0 ù é3 2ù é1 -1ù é7 3ù
ê2 3 ú ê1 2ú = ê9 10ú but ê1 2ú ê2 3 ú = ê5 5ú
ë û ë û ë û ë û ë û ë û

7.2 Unit and inverse matrices

Number one has a special property - multiplication by number one does not change other
numbers: 1 ´ y = y for any number y . Similar role in the matrix multiplication plays
unit matrix signed usually E : equality EA = A should hold for any matrix A .

Examples
é1 0ù é a b cù éa b cù
ê0 1ú êd e =
ë û ë f úû êd e
ë f úû

é1 0 0ù
éa b cù ê éa b cù
êd e ú 0 1 0úú = êd e
ë fû ê
ë f úû
êë0 0 1úû

We can easily see that the common form of unit matrix (it is also called identity matrix) is like
this:
 56

é1 0 0 0 . . 0ù
ê0 1 0 0 . . 0ú
ê0 0 1 0 . .. 0ú
E = ê. . . . . . .ú
ê. . . . . . .ú
ê0 0 0 0 . . 0ú
êë0 0 0 0 . . 1úû

These are square matrices having "ones" along the "main diagonal" and zeros elsewhere.
Unit matrices differ only in their size which must be proper for multiplication with respect to
the size of the multiplied matrix.

There is another basic property of multiplication of numbers - for any nonzero number x ,
there is exactly one number y so that xy = 1 holds. This number y is signed either 1/x or
x -1 . Analogous problem in matrix multiplication yields to very important concept of
inverse matrix. Let us have a square matrix A of size n ´ n . The inverse matrix to A is
matrix A -1 if equality A A -1 = E holds.

Example 42
é 2 1ù é0 -0,2ù = é1 0ù
ëê-5 0úû êë1 0,4 úû ëê0 1úû

é0 -0,2ù é 2 1ù = é1 0ù
êë1 0,4 úû êë-5 0úû êë0 1úû

As you can see in Example 42

if A A -1 = E , then A -1 A = E holds, too - in this special case the multiplication of

matrices is commutative.
Computing of inverse matrix is quite complicated and laborious matter, especially for
matrices of larger size. This job is done by the help of computers as a rule nowadays, but its
results can be very useful in practice, as we shall se in the next paragraph.
7.3 Application of inverse matrix to the solution of systems of linear equations
Let us have a system of n linear equations with n unknown variables xi like this:

a1,1 x1 + a1,2 x2 + . . . + a1,n xn = b1

a2,1 x1 + a2,2 x2 + . . . + a2,n xn = b2
. . . . . . . . . . . .
an,1 x1 + an,2 x2 + . . . + an,n xn = bn

where ai , j , bi are known constants. We shall transform this problem into matrix notation
like this:
 57

é a1,1 a1,2 ... a1,n ù é x1 ù éb1 ù

êa a 2 ,2 ... a 2 ,n ú ê ú ê ú
A = ê 2 ,1 ú X = ê x2 ú B = êb2 ú
ê . . ... . ú ... ...
êx ú êb ú
êëa n ,1 a n ,2 ... a n ,n úû ë nû ë nû

In this notation, our system of equations can be written down in following way:

é a1,1 a1,2 ... a1,n ù é x1 ù éb1 ù

êa a 2 ,2 ... a 2 ,n ú ê x 2 ú êb2 ú
ê 2 ,1 ú =
ê . . ... . ú ê ... ú ê ... ú
êëa n ,1 a n ,2 ... a n ,n úû êë x n úû êb ú
ë nû
or shortly
AX = B
(Try to multiply matrices AB and compare the result with matrix B.)
Now let as use inverse matrix A -1 and multiply both sides of our equation with it:
A -1 A X = A -1 B
but because A -1 A = E , the left side of the equation is EX = X . Hence, we get equation
X = A -1 B
and its meaning is following: if inverse matrix A -1 is known, then the solution of system of
linear equations reduces to simple multiplication of this matrix and the column matrix B .
This method is very useful especially in case that we need to solve similar systems of
equations again and again, where only values bi change.

Example 43
Dermatologists prescribe for some types of eczema the tincture containing two active
substances A,B and the solvent. Prescriptions on case to case basis vary in both required
concentrations CA and CB and in the total prescribed amount of the formulation. In the
pharmacy following components are available:
1) Solution of substance B with concentration CB1 = 5%
2) Solution of both substances with concentrations CA2 = 10% ; CB2 = 1%
3) pure solvent
For calculation of the required masses of particular components M1, M2, M3 with the
preassigned values of the total mass M and concentrations CA, CB it is necessary to solve the
system of equations
M1 + M 2 + M 3 = M
10 M 2 = CAM
5M 1 + 1M 2 = CB M
 58

As the particular inputs vary only in the right sides, solution using inverse matrices is suitable.
Matrix form of the system is C M = D ; in detail

é1 1 1ù éM1 ù é M ù
ê0 10 0úú êM ú = êC M ú
ê ê 2ú ê A ú
êë5 1 0 úû êë M 3 úû êëCB M úû

Calculation of the inverse matrix2 C-1 leads to the solution in form M = C-1 B, i.e.

éM1 ù é0 - 0.02 0.2 ù é M ù

ê M ú = ê0 0.1 0 úú êC M ú
ê 2ú ê ê A ú
êë M 3 úû êë1 - 0.08 - 0.2úû êëCB M úû
For the lab assistant who is in charge of preparation of the tincture we can transcribe it into
the following comprehensible form

M1 = - 0.02 C A M + 0.2 C B M
M2 = 0.1 C A M
M3 = M - 0.08C A M - 0.2C b M

Exercise 15
Try to sum the expressions for M1 , M2 and M3 and prove that M1 + M2 + M3 = M holds,
i.e. the third equation can be simplified as M3 = M – (M1 + M2).

2
Methods of computation of inverse matrix C-1 transcend the extent of this textbook. More about this will be
studied in the subject Applied computer technology using program Excel.