پديد آورنده :
پوربافراني ، ندا
عنوان :
پيش‌بيني انرژي تشكيل و گاف هومو - لوموي نانوخوشههاي سيليكون با استفاده از روش يادگيري ماشين
مقطع تحصيلي :
كارشناسي ارشد
گرايش تحصيلي :
حالت جامد (ماده چگال)
محل تحصيل :
اصفهان : دانشگاه صنعتي اصفهان
صفحه شمار :
چهارده ، 81ص. : مصور، جدول ، نمودار
استاد راهنما :
مجتبي اعلائي
استاد مشاور :
اسماعيل عبدالحسيني سارسري
توصيفگر ها :
روش‌هاي ابتدا به ساكن , تابع سطح انرژي پتانسيل , يادگيري ماشين , نانوخوشههاي سيليكون
استاد داور :
جواد هاشمي فر ، فرهاد شهبازي
تاريخ ورود اطلاعات :
1398/10/26
تاريخ ويرايش اطلاعات :
1398/11/02
چكيده انگليسي :
Prediction of the atomization energy and HOMO LUMOgap of Nano Silicon clusters using Machine Learning method Neda Pourbafrani n pourbafrani@ph iut ac ir Department of Physics Isfahan University of Technology Isfahan 84156 83111 Iran University Code IUT 77142 Degree M S c Language Farsi supervisor Dr Mojtaba Alaei m alaei at cc iut ac ir Abstract One of the fastest growing branches of material science is the prediction of material properties using alternative methods instead of ab initio methods such as density functional theory DFT because one of the main problems of computational ab initio methods is that with increasing the scale and size of the system the computational cost is rapidly increasing and requires heavy computing One of these alternatives is machine learning which significantly reduces these costs One of the valuable physical properties of a potential energy surface function is a multidimensional real value function that provides the potential energy of a system as a function of atomic positions The concept of potential energy surface is derived from the Born Oppenheimer approximation of quantum mechanics If the position of the atoms the charge of the nucleus and the total charge of a system are known one can calculate its potential energy with the electron Hamiltonian which has a heavy computational density functional theory In this project we predict the energy and HOMOa LUMOb gap of Nano Silicon clusters through machine learning methods by using two computational packages QMML and The RuNNer with potential Gaussian approximation method kernel tricks and artificial neural networks a b Highest Occupied Molecular Orbital Lowest Unoccupied Molecular OrbitalKeywords Ab initio methods Potential energy surface function Machine Learning NanoSilicon clusters
استاد راهنما :
مجتبي اعلائي
استاد مشاور :
اسماعيل عبدالحسيني سارسري
استاد داور :
جواد هاشمي فر ، فرهاد شهبازي
