Search In this Thesis
   Search In this Thesis  
العنوان
Applications of Artificial Neural Networks
in Physics /
المؤلف
Ali, Engy Hany Ali.
هيئة الاعداد
باحث / إنجي هاني علي علي
مشرف / محمود يس البكري
مشرف / رشا علي علي
مناقش / دعاء محمود حبشي
تاريخ النشر
2021.
عدد الصفحات
137p. :
اللغة
الإنجليزية
الدرجة
ماجستير
التخصص
الفيزياء والفلك (المتنوعة)
تاريخ الإجازة
1/1/2021
مكان الإجازة
جامعة أسيوط - كلية التربية - فزياء انجليزى
الفهرس
Only 14 pages are availabe for public view

from 137

from 137

Abstract

The term Artificial Inteligence refers to systems or devices that simulate human
intelligence to perform tasks that can improve themselves based on the information
they collect. Artificial intelligence manifests itself in a number of forms. Artificial
intelligence is more about the ability to think and analyze data than it is about a
specific shape or function.
Neural networks are a model of intelligence models and are known as (Nodes,
Neurons) to watch them with the property of experience, experimental information, to
make them, by adjusting the weights. It is analogous to the human brain in an industry
that gains knowledge by training and stores knowledge using forces within neurons
called synaptic weights. There is also a dynamic ring, which gives biologists the
opportunity to draw on these networks to understand the evolution of biological
phenomena.
The present work deals with a theoretical study of some parameters in solid state and
in p-p collisions at high energy in the framework of Artificial intelligence (AI) which
contains the model of Artificial Neural Network (ANN) for simulation these
parameters and also prediction the values of experiments that have not been made yet
with describing them by mathematical function obtained from obtained this model.
The first chapter : This chapter deals with a concept the Artificial intelligence ,
Philosophy of AI , Goals of AI , What Contributes to AI? , Programming without and
with AI , Applications of AI.
In the second chapter: We discussed the study of one model from Artificial
Intelligence (AI) which is the model of the Artificial Neural Network for studying
some photovoltaic characteristics of NiPc/p-Si and some characteristics of p-p
interactions .This chapter explains philosophy of AI , goals and techniques of AI ,
classification and task domain of AI, difference between Human and Machine
Intelligence, AI Contributions , applications of AI.
In the third chapter: Neural networks were designed for studying the model the
photovoltaic properties of Nickel–phthalocyanine (NiPc/p-Si) heterojunction. Eight
neural networks were trained to simulate the experimental data their constructions
were introduced in this chapter. The ANN predicted results at values of 298 K for
dark current–voltage characteristics and 373 K for variation of log I with log Vof
NiPc/p-Si heterojunction which measured experimentally introduce an excellent
accordance with the experimental data. Also, the predicted ANN results at 273 and
340 K for dark current–voltage characteristics and at 273 ,360 and 380 K for variation
x
of log I with log V of NiPc/p-Si heterojunction which are not measured previously
present logical results. The most important result presented is the mathematical
relation which describes the behavior of the photovoltaic properties of NiPc/p-Si
(organic/inorganic) heterojunction.
In the fourth chapter: Neural networks were designed for studying some parameters
of p-p interactions at high energy ( Inelastic p-p cross section, Particle pseudo
rapidity density, Transverse momentum distribution, Multiplicity dependence of
identified hadron production in pp, Invariant-mass distributions of Weaklydecaying
Strange Hadrons, Pseudo-Rapidity Distribution Differential yields of
particle/antiparticle production yields). Twenty neural networks were trained to
simulate the experimental data their constructions were introduced in this chapter.
1.Inelastic p-p cross section
Inelastic p-p cross section as a function of c.m. energy in the range √s ≈10
GeV–500 TeV from p-¯p (UA5 , E710 and CDF) and p-p (ALICE ATLAS,CMS,
TOTEM ) colliders, as well as the AUGER result at √s = 57 TeV derived from
cosmic-ray data. Neural Network is used to simulate these experimental data for
inelastic p-p cross section as a function of energy √s and compared to the
prediction of pythia.
2. Particle pseudo rapidity density
Neural Network was designed for studying Particle pseudo rapidity density for p-p
interactions. The c.m. energy evolution of the charged hadron pseudo rapidity
density at = 0 in the range √s = 10 GeV–800 TeV is inelastic data which
measured at Sp¯pS (UA5 )and LHC (ALICE, ATLAS and CMS) colliders.
Neural Network is used to simulate these experimental data for evolution of the
charged particle pseudo rapidity density at midrapidity, , as a
function of collision energy, √s, for inelastic p-p collisions and compared to the
prediction of pythia.
3. Transverse momentum distribution
Neural Network was designed for studying transverse momentum distribution for(
xi
p-p), (p-¯p )interactions Evolution of at midrapidity as a function of c.m.
energy √s. Data points show existing collider results. Neural Network is used to
simulate these experimental data for Evolution of at midrapidity as a function
of c.m. energy √s and compared to the prediction of pythia.
4. Multiplicity dependence of identified hadron production in pp
Neural Network was designed for studying Multiplicity dependence of charged
and neutral kaon yields obtained using mid-pseudo rapidity charged particle
multiplicities (| |< 0.5) and the charged particle multiplicities within the pseudo
rapidity. Data collected by ALICE in the LHC pp run of the year 2010 are used.
Neural Network is used to simulate these experimental data for multiplicity
dependence of charged and neutral kaon yields obtained using mid-pseudo
rapidity charged particle multiplicities (|ƞ |< 0.5).
5. Invariant-mass distributions of Weakly-decaying Strange Hadrons
Neural Network was designed for studying Invariant-mass distributions of
, Λ,
, k and kk decay candidates in selected ranges for multiplicity in pp
collisions at √ s = 7 TeV. Data collected by ALICE in the LHC pp run of the year
2010 are used. Neural Network is used to simulate these experimental data for
Invariant-mass distributions of
, Λ, , k and kk decay candidates in
selected ranges.
6. Pseudo-Rapidity Distribution
Neural Network was designed for studying Pseudo-Rapidity Distribution for p-p
interactions at √ s=2.36, the ISR(√ s =23.6 GeV ) to the Tevatron (CDF data, √ s =1.8
TeV). Increasing the energy results in an increase in multiplicity[, the Large Hadron
Collider (LHC) √s = 7 TeV and (√ s =8 TeV recorded by the ATLAS detector at the
LHC). Neural Network is used to simulate these experimental data for the pseudorapidity
distributions (d /dɳ ) of the created particles in p-p interactions at different
energies √s(23.6 ,53 ,200 ,900 ,1800. 2360,7000 ,8000 GeV)and predicted known
(2.36 TeV) and unkown energies (10 and 14 TeV) for the pseudo-rapidity
distributions (dn/dɳ ) of the created particles in p-p interactions at energies √s.
xii
7. Differential yields of particle/antiparticle production yields
Neural Network was designed for studing Differential yields of particle/antiparticle
production yields .The results are shown for a selection of event classes , indicated by
roman numbers in brackets ,with decreasing multiplicity. The data have been
performed ALICE detector at LHC -differential yields of ̅
̅
measured in |y|<0.5 for a selection of event classes with progressively
decreasing< /d > . Neural Network is used to simulate these experimental for
- differential yields of ̅ ̅
measured in |y|<0.5.
* The most important results obtained by applying the artificial neural network model
to simulate practical values related to some physical parameters related to solid state
physics and high energy physics:
1- A general mathematical equation describing the physical behavior of the subject of
study in the fields of solid state physics and high energy physics.
2-Using the artificial neural network model, high-precision simulation results were
obtained compared to the practical values of the physical parameters in the two fields
(solid state physics - high energies physics). Also, the accuracy of the high results was
proven when compared to the results obtained.
3- Using the artificial neural network model, process values included in the model
inputs were predicted (not practically measured) and high-precision results were
obtained.
Summarizing the above, the study demonstrated the high efficiency of the artificial
neural network model in simulating and predicting different physical parameters
related to solid state physics and high energy physics, as well as its ability to give a
general equation describing the physical behavior in the two cases.