Линейное программирование. Практикум. Учебное пособие для бакалавриата
.pdf
1919
Федеральное государственное образовательное бюджетное учреждение высшего образования «ФИНАНСОВЫЙ УНИВЕРСИТЕТ
ПРИ ПРАВИТЕЛЬСТВЕ РОССИЙСКОЙ ФЕДЕРАЦИИ» (Финансовый университет)
Кафедра математики
И.Г. Шандра
ЛИНЕЙНОЕ
ПРОГРАММИРОВАНИЕ
Практикум
Учебное пособие для бакалавриата экономических и математических вузов и факультетов
МОСКВА
2024
УДК 519.852 ББК 22.18
Ш20
Рецензенты:
Рылов А.А., канд. физ.-мат. наук, доцент Финансового университета при Правительстве Российской Федерации;
Попов В.Ю., доктор. физ.-мат. наук, профессор Национального исследовательского университета «Высшая школа экономики».
Шандра И.Г.
Ш20 Линейное программирование. Практикум: учебное пособие для бакалавриата / И.Г. Шандра. — М.: Прометей, 2024. — 112 с.
ISBN 978-5-00172-715-6
Учебное пособие включает в себя разделы: введение в линейное программирование, геометрический метод, симплекс метод, метод искусственных переменных, двойственность, транспортная задача. Пособие может быть использовано для изучения избранных глав курсов «Математика», «Линейная алгебра», «Линейное программирование», «Методы оптимизации», «Методы оптимальных решений» для подготовки бакалавров экономических и математических направлений, обучающихся на английском языке. Пособие может быть также интересно студентам магистратуры, аспирантам, научным работникам, учителям, преподавателям.
ISBN 978-5-00172-715-6 |
© Шандра И.Г., 2024 |
|
© Издательство «Прометей», 2024 |
1919
FEDERAL STATE EDUCATIONAL BUDGETARY INSTITUTION OF HIGHER EDUCATION “FINANCIAL UNIVERSITY UNDER THE GOVERNMENT OF THE RUSSIAN FEDERATION” (Financial University)
Chair of Mathematics
I.G. Shandra
LINEAR PROGRAMMING
Training Manual
Manual for undergraduate students
of economic and mathematical universities and faculties
Moscow
2024
УДК 519.852 ББК 22.18
Ш20
Reviewers:
Rylov A.A. — PhD of Math, Associate Professor of the Financial University under the Government of the Russian Federation;
Popov V.Yu. — Doctor of Physical and Mathematical Sciences, Professor of the National Research University “Higher School of Economics”.
Shandra I.G.
Ш20 Linear programming. Training manual: for undergraduate students of economic and mathematical universities and faculties. — Moscow: Prometheus, 2024. — 112 p.
ISBN 978-5-00172-715-6
The tutorial consists of six parts: introduction to linear programming, geometric method, simplex method, method of artificial variables, duality, transportation problem. The manual can be used to study selected chapters of the courses “Mathematics”, “Linear Algebra”, “Linear Programming”, “Optimization Methods”, “Methods of optimal solutions” for the preparation of bachelors of economic and mathematical fields studying in English. The manual may also be of interest to graduate students, postgraduates, researchers, teachers, professors.
ISBN 978-5-00172-715-6 |
© Shandra I.G., 2024 |
|
© Prometheus publishing house, 2024 |
CONTENT
1. |
Preface . . . . . . . . . . . . . . . . |
. . |
6 |
|
2. Introduction to linear programing. . . . . |
. . . |
. |
. 7 |
|
3. |
Geometrical method . . . . . . . . . . . |
. . . . 19 |
||
4. |
Simplex method . . . . . . . . . . . . |
. . . . |
31 |
|
5. |
Method of artificial variables. . . . . . . |
. . . |
. |
48 |
6. |
Duality . . . . . . . . . . . . . . . . |
. . |
57 |
|
7. |
Transportation Problem. . . . . . . . . |
. . . |
. |
73 |
Conclusion. . . . . . . . . . . . . . . . |
. . |
95 |
||
Answers for questions and exercises . . . . . |
. . . . 96 |
|||
Recommended literature. . . . . . . . . . |
. . . |
. 110 |
||
— 5 —
1. PREFACE

Linear programming as a science appeared due to the need to solve applied economic problems.
The founder of linear programming was the Soviet mathematician Leonid Kantorovich. In 1939, his first work was published on the application of linear programming to solve applied economic problems. In 1975, Leonid Kantorovich was awarded the Nobel Prize for the development of linear programming methods.
The American mathematician George Bernard Danzig also made a great contribution to the development of linear programming. In 1949, he developed an effective method for solving linear programming problems — the simplex method.
Thе manual consists of four parts; introduction to linear programming, geometric method, simplex method, artificial variables method. The manual can be used to read selected chapters of the courses “Mathematics”, “Linear Algebra”, “Linear Programming”, Optimization Methods”, “Methods of optimal solutions” for bachelor’s and master’s degrees in economics and mathematics. The manual may also be of interest to graduate students, postgraduates, researchers, teachers, university professors.
— 6 —
2. INTRODUCTION TO LINEAR
PROGRAMMING
Linearprogrammingisabranchofappliedmathematics that deals with solving problems of maximization or minimization of a linear function subject to linear constraints (it means that constraints are linear equations or linear inequalities).
Here is an example of a linear programming problem in mathematical form.
Example.
f =2x1 −x2 +2x3 +3 → max
Subject to
x1 +6x2 −2x3 = 4,5x1 +3x2 ≥5,
x1 ≥ 0,x2 ≥ 0,x3 ≥ 0.
Now let us consider important economic problems that lead to linear programming problems.
|
Example. (The diet problem). |
|
|
Suppose that a diet consists of m different types |
|
of |
foods: F1, ,Fm that supply n |
nutrients: N1, ,Nn . |
Let |
aij be the quantity of nutrient |
Ni contained in one |
unit of food Fj . A diet is healthy if it has daily at least q1, ,qn quantities of the nutrients N1, ,Nn respectively. To compose such a diet, we need to buy quantities x1, ,xm of the foods having prices p1, , pm . The problem is to supply the required nutrients at minimum cost.
— 7 —
2. Introduction to LinearProgramming
f = p1x1 + + pmxm → mina11x1 + +a1mxm ≥ q1,
,
an1x1 + +anmxm ≥ qn,
x1 ≥ 0, ,xm ≥ 0.
Or in the compact matrix form
f = PT X → min
AX ≥ Q,
X ≥ 0.
Example. (The planning production problem).
Suppose a firm produces m types of goods: G1, ,Gm
having the prices p1, , pm and |
uses |
in |
the production |
n types of resources: R1, ,Rn |
Let |
aij |
be the quantity |
of resource Ri used in the production of one unit of good Gj and q1, ,qn be the supplies of resources. The problem is to find the outransportation problem uts of productions (the optimal plan) x1, ,xm maximizing the total income.
f = p1x1 + + pmxm → maxa11x1 + +a1mxm ≤ q1,
,
an1x1 + +anmxm ≤ qn,
x1 ≥ 0, ,xm ≥ 0.
Or in the compact matrix form
f = PT X → max
AX ≤ Q,
X ≥ 0.
— 8 —
2. Introduction to LinearProgramming
Example. (The transportation problem). Let there be m producers of some good: A1, , Am (points of departure)
and n consumers B1, ,Bn (points of destination). It is given for each producer the supply ai ≥ 0 (i =1, ,m) and for each customer the demand bj ≥ 0 (j =1, ,n) . It is known the unit cost cij ≥ 0 of transportation of the good from the i-th producer to the j-th customer. It is required to find the volume of transportation xij from the i-th producer to the j-th customer minimizing the total cost of transportation.
Definition. A linear programming problem is the optimizing problem of the following type:
f = c1x1 + +cmxm +c0 → max (min)
m n |
|
|
|
f = ∑∑cijxij → min |
|
||
i=1 j=1 |
|
|
|
m |
|
j =1, ,n, |
|
∑xij = bj, |
|
||
i=1 |
|
|
|
n |
|
i =1, ,m, |
|
∑xij = ai, |
|
||
j=1 |
|
|
|
x ≥ 0. |
|
|
|
ij |
|
|
|
|
|
|
|
Subject to |
+ +a1mxm ≤ b1, |
|
|
a11x1 |
|
||
, |
(1) |
||
|
+ +anmxm ≤ bn, |
|
|
an1x1 |
|
||
d11x1 + +d1mxm = q1, |
|
||
, |
(2) |
||
|
+ +dkmxm = qk, |
|
|
dk1x1 |
|
||
— 9 —
2. Introduction to LinearProgramming
x1 ≥ 0, ,xm ≥ 0 . |
(3) |
Or in the compact matrix form
f = CT X +C0 → max (min)
AX ≤ B,
DX = Q,
X ≥ 0.
The linear function f to be maximized or minimized is called the objective function.
Remark. If we have the problem of finding the maximum of the objective function f , then the problem for the objective function g = −f under the same constraints will be the minimization problem and vice versa.
A solution X* = (x1*,...,xn* )T of the linear programming problem is called the optimal solution (vector).
The constraints (1), (2) are called the main constraints and the constraints (3) are called the non-negativity constraints.
A vector X = (x1,...,xn )T satisfying the constraints of the linear programming problem is called a feasible solution (vector).
The set of all feasible solutions is called the feasible set
(region).
A linear programming problem is said to be feasible if the feasible set is not empty otherwise it is called infeasible.
Definition. A set M is said to be convex if for any x,y M and any t [0;1] it follows that tx +(1−t)y M .
From the geometrical point of view of geometry convexity of set means that if two points belong to the set then the segment connecting these points also belongs to the set.
— 10 —
