Авторы

  • Muzaffar Sulaymonovich Azizov
  • Odina Isroiljon qizi Ismoiljonova

DOI:

https://doi.org/10.71337/inlibrary.uz.esiiw.125373

Ключевые слова:

Gradient method steepest descent iterative method symmetric matrix positive definiteness functional optimization Python algorithm error vector.

Аннотация

This article discusses the steepest descent algorithm of the gradient method used to solve systems of linear algebraic equations. The method for finding the solution by minimizing a functional is explained step-by-step. Theoretical foundations 
based on gradient and error vectors are presented. A practical example involving a system of four equations is solved using Python, and the convergence rate and accuracy of the algorithm are demonstrated. The results show that the gradient method is a 
simple and efficient computational tool suitable for solving large-scale linear systems. 


background image

ОБРАЗОВАНИЕ НАУКА И ИННОВАЦИОННЫЕ ИДЕИ В МИРЕ

https://scientific-jl.org/obr

Выпуск журнала №-69

Часть–4_ Мая –2025

444

2181-3187

INVERSE BOUNDARY VALUE PROBLEM

Muzaffar Sulaymonovich Azizov

Senior Lecturer at the Department o

f Mathematical Analysis and Differential Equations,

Fergana State University

PhD in Physics and Mathematics

E-mail:

muzaffar.azizov.1988@mail.ru

Odina Isroiljon qizi Ismoiljonova

3rd-year student of the Applied Mathematics program,

Group 22-08, Fergana State University

E-mail:

ismoiljonovaodina88@gmail.com

Gulnoza Abrorjon qizi Ibrohimova

3rd-year student of the Applied Mathematics program,

Group 22-09, Fergana State University

E-mail:

jabborovagulniza2004@gmail.com

Annotation:

This article discusses the steepest descent algorithm of the gradient

method used to solve systems of linear algebraic equations. The method for finding the

solution by minimizing a functional is explained step-by-step. Theoretical foundations

based on gradient and error vectors are presented. A practical example involving a

system of four equations is solved using Python, and the convergence rate and accuracy

of the algorithm are demonstrated. The results show that the gradient method is a

simple and efficient computational tool suitable for solving large-scale linear systems.

Key words:

Gradient method, steepest descent, iterative method, symmetric

matrix, positive definiteness, functional, optimization, Python, algorithm, error vector.


background image

ОБРАЗОВАНИЕ НАУКА И ИННОВАЦИОННЫЕ ИДЕИ В МИРЕ

https://scientific-jl.org/obr

Выпуск журнала №-69

Часть–4_ Мая –2025

445

2181-3187

Аннотация:

В статье рассматривается обратная задача определения

граничного условия для уравнения теплопроводности. Неизвестная функция,

зависящая от времени, определяется на основе дополнительной информации о

решении. Существование и единственность решения доказываются с помощью

теорем 1 и 2. Кроме того, показано, что задача может быть сведена к

интегральному уравнению типа Вольтерра.

Ключевые слова:

обратная задача, уравнение теплопроводности,

граничное условие, интегральное уравнение, единственность, уравнение

Вольтерра

Annotatsiya:

Ushbu maqolada issiqlik tarqalish tenglamasi uchun chegaraviy

shartni aniqlashga oid teskari masala ko‘rib chiqiladi. Noma’lum bo‘lgan vaqtga

bog‘liq funksiyani aniqlash uchun yechim haqidagi qo‘shimcha ma’lumotlardan

foydalaniladi. Masalaning yechimining mavjudligi va yagonaligi 1- va 2-teoremalar

orqali asoslanadi. Bundan tashqari, bu masala Volterra turidagi integral tenglama orqali

yechim topishga keltirilishi ham ko‘rsatib beriladi.

Kalit so’zlar:

teskari masala, issiqlik tarqalish tenglamasi, chegaraviy shart,

integral tenglama, yagonalik, Volterra tenglamasi

Introduction:

Achieving fast and accurate solutions to systems of linear algebraic equations is

one of the key directions of modern computational mathematics. In particular, for

systems with symmetric and positive definite coefficient matrices, gradient-based

methods—especially the steepest descent algorithm—are highly effective. This

method operates on an iterative approach, moving in the direction of the greatest

decrease of the objective function at each step. This article presents the theoretical

foundations of the gradient method, the steps of the algorithm, and a practical example

implemented using the Python programming language.


background image

ОБРАЗОВАНИЕ НАУКА И ИННОВАЦИОННЫЕ ИДЕИ В МИРЕ

https://scientific-jl.org/obr

Выпуск журнала №-69

Часть–4_ Мая –2025

446

2181-3187

Determination of Boundary Conditions.

We consider an inverse problem

that involves determining a time-dependent function included in the boundary

condition, using additional information about the solution of the boundary value

problem for the heat conduction equation. Let us assume that

( , )

u x t

the function

2

, 0

, 0

,

t

xx

u

a u

x

l

t

T

=

 

 

(1)

(0, )

( ),

u

t

t

=

0

,

t

T

 

(2)

( , )

( ),

x

u l t

t

=

0

,

t

T

 

(3)

( , 0)

0,

u x

=

0

,

x

l

 

(4)

boundary issues of the solution , let it be.

Assume let,

( )

t

given the function,

( )

t

of unknown function, and (1)-(4)

of problem solution about

( , )

( ),

u l t

g t

=

0

,

t

T

 

(5)

additional information is,

( )

t

the function to determine required will be, this

earth

( )

g t

- given function.

Let us consider the issue of the uniqueness of the solution to the given inverse

problem

.

The uniqueness of the solution to this problem has been studied in a more

general form by a number of authors. Let us present one of the results obtained in this

direction, as provided in [51], and apply it to the case of the heat conduction equation

(1) with constant coefficients.

We introduce the following notation:

{( , ) : 0

, 0

}.

lT

Q

x t

x

l

t

T

=

 

 


background image

ОБРАЗОВАНИЕ НАУКА И ИННОВАЦИОННЫЕ ИДЕИ В МИРЕ

https://scientific-jl.org/obr

Выпуск журнала №-69

Часть–4_ Мая –2025

447

2181-3187

1-teorema.

Assume let me, the function

2,1

( , )

(

)

lT

u x t

C

Q

and (1) equations of

the

lT

Q

area will build. It without , you

0

t

T

 

for

( , )

( , )

0

x

u l t

u l t

=

=

be,

lT

Q

at

( , )

0

u x t

=

will.

This teorema from given

( )

t

and

( )

g t

the function for (1)-(5) from

( )

t

the

function identifying issues of the singleness come out can. Fact is,

2,1

1

2

( , ),

( , )

(

)

lT

u x t u x t

C

Q

lT

Q

in the area of (1) the equation may build and

1

1

(0, )

( ),

u

t

t

=

2

2

(0, )

( ),

u

t

t

=

0

,

t

T

 

1

2

( , )

( , )

( ),

u l t

u l t

g t

=

=

1

2

( , )

( , )

( ),

du

du

l t

l t

t

dx

dx

=

=

0

t

T

 

which is.

Now

1

2

( , )

( , )

( , )

u x t

u x t

u x t

=

the function we do not look at. This function is 1-

teorema terms and conditions will be content. Therefore,

lT

Q

in the field ,

( , )

0

u x t

=

ie

[0, ]

t

T

for

1

2

( )

( ).

t

t

=

Attention give, 1-teorema in

( , )

u x t

the function for the initial conditions, it is

not given.

(1)-(4) boundary issues for other reverse the issue , we do not look at.

Assume let,

( )

t

given the function,

( )

t

unknown following

0

( , )

( ),

u x t

g t

=

0

t

T

 

(6)

Wholesale of view (1)-(4) of problem solution about additional information

known if it is,

( )

t

the function to identify necessary, this earth on

( )

g t

a given

function,

0

[0, ].

x

l

The uniqueness of the solution to this inverse problem can also be

established using Theorem 1.


background image

ОБРАЗОВАНИЕ НАУКА И ИННОВАЦИОННЫЕ ИДЕИ В МИРЕ

https://scientific-jl.org/obr

Выпуск журнала №-69

Часть–4_ Мая –2025

448

2181-3187

2-teorema.

If

2,1

1

2

( , ),

( , )

(

)

lT

u x t u x t

C

Q

the function of the

lT

Q

area in (1) in

the equation and (3), (4), (6) terms and conditions does it build,

[0, ]

t

T

for

(0, )

( ),

1, 2,

i

i

u

t

t i

=

=

is. It without

[0, ]

t

T

though

1

2

( )

( )

t

t

=

it will be.

Proof.

We consider the function

1

2

( , )

( , )

( , )

u x t

u x t

u x t

=

where

( , )

u x t

2

0

,

, 0

,

t

xx

u

a u

x

x

l

t

T

=

 

 

0

( , )

0,

u x t

=

0

,

t

T

 

( , )

0,

x

u l t

=

0

,

t

T

 

( , 0)

0,

u x

=

0

,

,

x

x

l

 

boundary of the issue the solution is. Optional

0

0

[ , ],

x

x l

[0, ]

t

T

for

( , )

0

u x t

=

this is that we show. For this, we multiply the equation by a function

( , )

u x t

integrate, and

0

0

2

0

0

( , ) ( , )

( , ) ( , )

l t

l t

t

xx

x

x

u

u

d d

a

u

u

d d

 

   

 

   

=

 

 

ensure you will. Integrated with many pieces in the calculation, the initial and

boundary conditions we will use, the result

[0, ]

t

T

for the following equality able

you will be:

0

0

2

2

2

0

1

( ( , ))

(

( , ))

2

l

l t

x

x

x

u

d

a

u

d d

 

 

 

+

 

That come out,

0

[ , ],

x

x l

[0, ]

t

T

to

( , )

0

u x t

=

. It without

0

( , )

0.

u x t

=

So done,

0

[0,

],

x

x

[0, ]

t

T

when

( , )

u x t

the function (1) in the equation and

[0, ]

t

T

for

0

0

( , )

( , )

0

u x t

u x t

=

=

terms and conditions will build. We apply Theorem 1 on the

rectangle

0

0

, 0

x

x

t

T

 

 

, for

[0, ]

t

T

, and thus

1

2

(0, )

(0, )

u

t

u

t

=

or

1

2

( )

( )

t

t

=

ollows. The proof of Theorem 2 is completed..


background image

ОБРАЗОВАНИЕ НАУКА И ИННОВАЦИОННЫЕ ИДЕИ В МИРЕ

https://scientific-jl.org/obr

Выпуск журнала №-69

Часть–4_ Мая –2025

449

2181-3187

Let us present an example of reducing the problem of determining a

boundary condition to the solution of a first-kind integral equation.

We consider the boundary value problem for the heat conduction equation on the

half-line:

, 0

, 0

,

t

xx

u

u

x

t

T

=

    

(7)

(0, )

( ),

u

t

t

=

0

,

t

T

 

(8)

( , 0)

0,

u x

=

0

.

x

  

(9)

(7)-(9) problem solution about

0

( , )

( ),

u x t

g t

=

0

,

t

T

 

0

0,

x

(10)

Additional information is given,

( )

t

the function to identify required is.

(7)-(9) the issue of the solution

2

3 2

0

( , )

exp

( )

4(

)

2

(

)

t

x

x

u x t

d

t

t

  

=

has the following form. Thus, in this case, the inverse problem

0

( , ) ( )

( ),

t

K t

d

g t

   

=

0

,

t

T

 

It is reduced to a Volterra integral equation of the first kind, where the kernel

is...

2

0

0

3 2

( , )

exp

.

4(

)

2

(

)

x

x

K t

t

t

=

General Conclusion:

In this paper, an inverse problem related to the determination of a boundary

condition for the heat conduction equation was investigated. The problem involved


background image

ОБРАЗОВАНИЕ НАУКА И ИННОВАЦИОННЫЕ ИДЕИ В МИРЕ

https://scientific-jl.org/obr

Выпуск журнала №-69

Часть–4_ Мая –2025

450

2181-3187

determining an unknown time-dependent function using additional information about

the solution. Based on Theorems 1 and 2, the uniqueness of the solution was proven.

Furthermore, a practical method for obtaining the solution was demonstrated by

reducing

the

problem

to

a

Volterra-type

integral

equation.

The results of this study can be applied in modeling and controlling physical processes

and contribute to the theoretical and practical understanding of inverse problems.

References:

1.

Lavrentyev, M.M., Romanov, V.G., Yanyushkin, S.P. –

Obratnye

zadachi matematicheskoy fiziki

. Moscow: Nauka, 1986.

2.

Tikhonov, A.N., Arsenin, V.Y. –

Reshenie nekorrektnykh zadach

.

Moscow: Nauka, 1979.

3.

Isakov, V. –

Inverse Problems for Partial Differential Equations

.

Springer, 2006.

4.

Cannon, J.M. –

The One-Dimensional Heat Equation

. Addison-

Wesley, 1984.

5.

Aliev, N.A. –

Introduction to the Theory of Inverse Problems

.

Tashkent: Fan, 1998.

6.

Mukhamedov, A.K. –

Equations of Mathematical Physics and Their

Inverse Problems

. Fergana: FSU Publishing, 2022.

7.

Rasskazov, A.V. –

Theoretical Foundations of Methods for Solving

Inverse Problems

. Novosibirsk, 2001.

Библиографические ссылки

Lavrentyev, M.M., Romanov, V.G., Yanyushkin, S.P. – Obratnye

zadachi matematicheskoy fiziki. Moscow: Nauka, 1986.

Tikhonov, A.N., Arsenin, V.Y. – Reshenie nekorrektnykh zadach.

Moscow: Nauka, 1979.

Isakov, V. – Inverse Problems for Partial Differential Equations.

Springer, 2006.

Wesley, 1984.

Cannon, J.M. – The One-Dimensional Heat Equation. Addison

Aliev, N.A. – Introduction to the Theory of Inverse Problems.

Tashkent: Fan, 1998.

Mukhamedov, A.K. – Equations of Mathematical Physics and Their

Inverse Problems. Fergana: FSU Publishing, 2022.

Rasskazov, A.V. – Theoretical Foundations of Methods for Solving

Inverse Problems. Novosibirsk, 2001.

Наиболее читаемые статьи этого автора (авторов)