teaching:methcalchim:start [Didier Villers, UMONS

Cette page est en lecture seule. Vous pouvez afficher le texte source, mais ne pourrez pas le modifier. Contactez votre administrateur si vous pensez qu'il s'agit d'une erreur.
====== Calculation methods applied to chemistry ======
===== Synopsis (english) =====
====  Mathematical prerequisites====

==== Programming bases and tools ====
  * Python programming language
    * [[http://www.learnpython.org/|LearnPython.org]] interactive tutorial with code execution
    * [[https://www.datacamp.com/courses/intro-to-python-for-data-science|DataCamp free course "Intro to Python for Data Science"]]
    * [[https://www.sololearn.com/Course/Python/|Python 3 Tutorial]], interactive, with [[https://code.sololearn.com/#py|code use in web browser]]
    * MOOCs (massive open online courses) :
      * [[https://www.coursera.org/course/interactivepython|An Introduction to Interactive Programming in Python (Beginners)]] (Coursera)
      * [[https://www.edx.org/course/introduction-computer-science-mitx-6-00-1x-9|Introduction to Computer Science and Programming Using Python]] (edX)
  * [[wp>Anaconda_(Python_distribution)|Anaconda Python distribution]] (64 bits, with Python 3.x) : [[https://www.anaconda.com/download/]]
    * Includes these tools :
      * Jupyter notebook (interactive web-based environment)
      * qtconsole (high level Python console with graphics & colors)
      * spyder (powerful Python IDE)
    * includes lot of Python libraries : matplotlib, numpy, scipy, pandas,...
    * package management system (conda), virtual environments
    * Anaconda Navigator includes extensive documentation (on anaconda website and dedicated websites)
  * GNU/Linux OS (preferred)

  * Jupyter introductions, tutorials, ...
    * [[https://medium.com/@denkasyanov/jupyter-notebook-tutorial-9c0ffa5ae9a1#.ufaor34wp|Jupyter Notebook Tutorial]], par Den Kasyanov (Medium)
    * [[https://jupyter.readthedocs.io/en/latest/|Jupyter official documentation]]
    * [[https://www.datacamp.com/community/blog/jupyter-notebook-cheat-sheet|Jupyter DataCamp Cheat Sheet]]

<note tip>The Microsoft Azure Notebooks environment can be used to execute sample codes, using a professionnal,  personnal or student login/account (i.e. student login from UMONS). The following public sample notebooks are then available to test Jupyter text and coding features :
  * [[https://notebooks.azure.com/linusable/libraries/samples-public]]
    * //e.g. :// [[https://notebooks.azure.com/linusable/libraries/samples-public/html/notebooks/second_degree_polynomial_and_roots-01.ipynb]] (Second degree polynomial and roots)
</note>

  * Python scientific libraries (official websites including tutorials and documentation)
   * //cf.// [[teaching:progappchim:start#les_librairies_scientifiques_essentielles|this]] (french)
   * [[http://matplotlib.org/|Matplotlib]] (scientific graphs)
  * [[http://www.numpy.org/|NumPy]] (array manipulations, linear algebra, Fourier transforms, random numbers,...
  * [[http://scipy.org/scipylib/index.html|SciPy]] numerical methods (integrations, ODE, PDE,...)
  * [[http://sympy.org/en/index.html|SymPy]] symbolis maths
  * [[http://pandas.pydata.org/|Pandas]], data analysis
  * Pylab, combine Matplotlib, NumPy and SciPy
  * [[http://scikit-learn.org/|Scikit-learn]], machine learning

==== Fundamental numerical methods ====
  * [[system_of_linear_equations|Systems of linear equations]]
    * Diagonalisation and triangularisation
    * LU decomposition : factorization in triangular matrices
  * [[root-finding_algorithm|Root findings :  equations f(x) = 0]]
    * Polynomial equations
    * Dichotomy
    * Secant method, Regula falsi
    * Newton-Raphson method
  * [[numerical_integration|Numerical intégration]] (integrals)
    * Simpson method and gaussian quadratures

<note tip>
Learning outcomes :
  * Systems of linear equations
    * failing of the theoretical way to solve a linear system using determinant and cofactors (np complexity)
    * triangularisation and diagonalisation principles : algorithm and complexity
    * "divide by zero" errors and pivot solutions
    * extension towards the matrix inversion
    * lower-upper LU decomposition and complexity (N³ for the decomposition step and N² for substitution step). How to solve systems with varying independant vectors
    * special matrix require special algorithms : tridiagonal matrix algorithm (Thomas algorithm)
  * Root findings
    * Bisection method (dichotomy) : simple and robust algorithm, invariant loop, slow convergence
    * iterative transformation x = f(x), convergence and divergence situations
    * secant and regula falsi methods, Convergence Criterion of the Fixed Point Method
    * Newton-Raphson method (use of derivatives), quadratic convergence, failure, tolerance and stop condition
    * Van Wijngaardeb-Dekker-Brent method ("black box" in numerical packages)
    * Roots of polynomials and Bairstow's method
  * Numerical intégration
    * Equally Spaced methods (trapezoidal, Simpson), accuracy, error dependance,...
    * Gaussian Quadratures and orthogonal polynomials (special integrals, scale transformations,  error estimates,...)
</note>

==== Classical numerical methods ====
=== Ordinary_differential_equations (ODE) ===

[[numerical_methods_for_ordinary_differential_equations|Numerical solutions of ODE]]
  * principe de discrétisation, méthode d'Euler
  * Améliorations et méthodes de Runge-Kutta
  * Runge-Kutta d'ordre 4
  * Contrôle du pas d'intégration
  * Méthodes predictor-corrector
  * Méthodes d'extrapolation (Richardson, Burlish-Stoer)
  * applications :
    * équations de cinétique chimique
    * Équation logistique
      * [[wp>fr:Modèle_de_Verhulst|Modèle de Verhulst]], [[wp>fr:Dynamique_des_populations|dynamique des populations]] et non-linéarité
      * Bifurcation, doublements de périodes et transition vers le chaos
    * Réactions chimiques oscillantes : Belousov-Zhabotinsky, Brusselator, Oregonator
    * Modèle proie-prédateur
    * Attracteur étrange, modèle atmosphérique de Lorenz

=== Partial_differential_equations (PDE) ===
[[partial_differential_equation|Numerical solutions of PDE]]
  * Domaine d'application des équations : équation de diffusion, équation d'ondes, équations de Navier-Stokes
  * Types de traitements numériques
  * Différences finies et problèmes de diffusion
  * Schémas classiques de différences finies
    * Résolutions stationnaires
    * Résolutions dépendantes du temps
  * Méthodes explicites, critère de (ou d'in)stabilités et méthodes implicites


=== Eigenvalues and eigenvectors ===
[[eigenvalues_and_eigenvectors|Eigenvalues and eigenvectors]]

applications à des problèmes de relaxation et de population, analyse de modes normaux de vibration, PCA (principal component analysis),...

=== Non-linear systems of equations ===
  * Newton-Raphson method


=== Linear and non-linear least squares approximations ===
  * Application to deconvolution (Levenberg–Marquardt algorithm)
  * Reférences :
    * [[https://pubs.acs.org/doi/abs/10.1021/acs.jchemed.8b00649|An Open-Source, Cross-Platform Resource for Nonlinear Least-Squares Curve Fitting]] Andreas Möglich, J. Chem. Educ., 2018, 95 (12), pp 2273–2278 DOI: 10.1021/acs.jchemed.8b00649


=== Chebyshev approximation ===
+ discussion of some approximations like [[wp>Bhaskara_I's_sine_approximation_formula|Bhaskara I's sine approximation formula]]
  * [[https://twitter.com/fermatslibrary/status/1267450081151782913]]
  * [[https://medium.com/@mathcube7/chebyshev-interpolation-with-python-2f2e89bb7c30|Chebyshev Interpolation With Python]] Mathcube, Medium, 03/02/2023

=== Molecules modelisation and visualization ===

=== Minimization ===
Conformational problems

----

===== Additional subjects =====
  * Bioinformatics and related algorithm (biochemistry, mass spectrometry,...)
  * Chemistry
    * quantum calculations, optimization, molecular mechanics
    * visualization, virtual reality
    * chemical informations on structures and reactions
  * Data science, statistics (Python modules : Scipy, Pandas,...)
    * Time series analysis
    * Machine learning (Scikit-learn,...)
  * Data visualization
    * boxplot, 3D, animations, graphs,...
  * Sensors and interfaces, Arduino, Raspberry Pi, IoT
  * Simulations
    * Agent base modelling and complex systems
    * cellular automaton
    * Simpy, active matter simulations...
  * Digital image processing, image recognition
    * particle tracking,...
    * Voronoi diagrams, Delaunay triangulation,... ([[https://towardsdatascience.com/the-fascinating-world-of-voronoi-diagrams-da8fc700fa1b|ref1]])


===== References =====
  * Gradient descent optimization
    * [[https://machinelearningmastery.com/gradient-descent-optimization-from-scratch/|How to Implement Gradient Descent Optimization from Scratch]] By Jason Brownlee on April 16, 2021
    * [[https://towardsdatascience.com/implementing-gradient-descent-in-python-from-scratch-760a8556c31f|Implementing Gradient Descent in Python from Scratch]] Vatsal Sheth, Medium, 18/02/2022
  * Bioinformatics
    * [[http://biopython.org/|Biopython]]
    * [[http://rosalind.info/problems/locations/|Rosalind.info]], learning bioinformatics by problem solutions
  * Machine Learning
    * Scikit-learn
      * [[https://towardsdatascience.com/linear-regression-in-6-lines-of-python-5e1d0cd05b8d|Linear Regression in 6 lines of Python]] (using scikit-learn)
    * [[https://datafloq.com/read/12-algorithms-every-data-scientist-should-know/2024|12 Algorithms Every Data Scientist Should Know]]
  * Deep Learning
    * TensorFlow
    * [[https://www.datacamp.com/community/tutorials/deep-learning-python#gs.PPrcRR4|Keras Tutorial: Deep Learning in Python]]
  * Arduino
    * [[http://pubs.acs.org/doi/abs/10.1021/acs.jchemed.6b00639|Fabrication of an Economical Arduino-Based Uniaxial Tensile Tester]], Julien H. Arrizabalaga, Aaron D. Simmons, and Matthias U. Nollert, J. Chem. Educ., 2017, 94 (4), pp 530–533 DOI: 10.1021/acs.jchemed.6b00639
  * PCA (principal component analysis)
    * [[http://pubs.acs.org/doi/abs/10.1021/acs.jchemed.6b00550|Learning Principal Component Analysis by Using Data from Air Quality Networks]], Luis Vicente Pérez-Arribas, María Eugenia León-González, and Noelia Rosales-Conrado, ,J. Chem. Educ., 2017, 94 (4), pp 458–464 DOI: 10.1021/acs.jchemed.6b00550
    * [[http://benalexkeen.com/principle-component-analysis-in-python/|Principle Component Analysis in Python]]
  * K-Means
    * [[http://benalexkeen.com/k-means-clustering-in-python/|K-means Clustering in Python]]
    * [[https://medium.com/thrive-in-ai/k-means-clustering-in-python-9f00b3d7abc3|K Means Clustering in Python]] Rohit Raj, Medium, 02/03/2022
    * [[https://medium.com/thrive-in-ai/classification-algorithms-in-python-5f58a7a27b88|Classification Algorithms in Python]] Rohit Raj, Medium, 15/03/2022
  * Applications (suggestions, examples,...)
    * [[https://scicomp.stackexchange.com/questions/28195/2d-ising-model-in-python|2D Ising Model in Python]]
  * chemistry
    * misc docs :
      * [[https://pdfs.semanticscholar.org/3161/89e419212794e74bf83442f65de96b5320ba.pdf|Extraction of chemical structures and reactions from the literature]]
    * [[https://medium.com/nerd-for-tech/animating-schrodinger-wave-function-%CF%88-of-a-particle-using-python-with-full-code-5ad9e4852906|Animating Schrodinger Wave Function(ψ) of a Particle Using Python (with full code) - Solving Particle in a Box Using Crank-Nicolson Method]], Kowshik chilamkurthy, Medium, 02/03/2021
  * Blog articles
    * [[https://py.plainenglish.io/linear-regression-in-plain-python-3b20bb56b31d|Simple Linear Regression Explained With Python - Explained in details which are easy to understand]] Arjan de Haan, Medium, Jan 23 (+ [[https://github.com/Vepnar/Script-Collection/blob/master/linear-regression/Linear%20regression.ipynb|jupyter notebook]])
    * simpy : [[https://towardsdatascience.com/simulate-real-life-events-in-python-using-simpy-e6d9152a102f|Simulate Real-life Events in Python Using SimPy]] by Khuyen Tran, May, 2021, Medium, Towards Data Science
    * Active matter simulations ([[https://arxiv.org/pdf/cond-mat/0611743.pdf|Vicsek]], 1995) :
      * [[https://medium.com/swlh/create-your-own-active-matter-simulation-with-python-76fce4a53b6f|Create Your Own Active Matter Simulation (With Python)]] Philip Mocz, The Startup, Medium 2021 + [[https://github.com/pmocz/activematter-python]]
    * Fourier transforms : [[https://towardsdatascience.com/fourier-transforms-an-intuitive-visualisation-ba186c7380ee|Fourier Transforms: An Intuitive Visualisation - An intuitive visualization of discrete Fourier transforms applied to simple time-series data]] Diego Unzueta, Towards Data Science, Medium, 17/09/2021
    * Monte-Carlo simulations : [[Monte Carlo Simulation — a practical guide - A versatile method for parameters estimation. Exemplary implementation in Python programming language]] Robert Kwiatkowski, Medium, 31/01/2022
    * régression logistique : [[https://mlu-explain.github.io/logistic-regression/]]
    * [[https://towardsdatascience.com/cubic-splines-the-ultimate-regression-model-bd51a9cf396d|Cubic Splines: The Ultimate Regression Model - Why cubic splines are the best regression model out there]] Brendan Artley, Medium, 27/07/2022








==== Books ====
  * [[http://www.springer.com/statistics/computational+statistics/book/978-3-642-28069-6|Solving Differential Equations in R]], chez Springer, et  en version électronique sur [[http://www.springerlink.com/content/978-3-642-28070-2#section=1080286&page=1|SpringerLink]]
  * [[http://www.cambridge.org/us/academic/subjects/engineering/engineering-mathematics-and-programming/numerical-methods-engineering-python-3-3rd-edition|Numerical Methods in Engineering with Python 3]] 3rd Edition, Jaan Kiusalaas, 2013, Cambridge University Press, isbn: 9781107033856
  * ...

==== Jupyter notebooks ====
  * [[https://github.com/mandli/intro-numerical-methods|Jupyter notebooks and other materials developed for the Columbia course APMA 4300]]
  * [[http://nbviewer.jupyter.org/url/norvig.com/ipython/Probability.ipynb|A Concrete Introduction to Probability (using Python)]]

==== MOOCs ====
  * [[https://openedx.seas.gwu.edu/courses/course-v1:MAE+MAE6286+2017/about|Practical Numerical Methods with Python]]

==== Miscellaneous ====
  * [[https://github.com/numerical-mooc/numerical-mooc|Practical Numerical Methods with Python]]
  * [[https://github.com/cfgnunes/numerical-methods-python]] (?)
  * [[https://medium.com/@mjspeck/presenting-code-using-jupyter-notebook-slides-a8a3c3b59d67|Presenting Code Using Jupyter Notebook Slides]]
  * [[https://choosealicense.com]]
  * [[https://creativecommons.org/choose/?lang=en]]
  * Published applications
    * [[http://pubs.acs.org/doi/10.1021/acs.jchemed.7b00395|Introduction to Stochastic Simulations for Chemical and Physical Processes: Principles and Applications]] Charles J. Weiss, Journal of Chemical Education 2017 94 (12), 1904-1910 DOI: 10.1021/acs.jchemed.7b00395
  * [[https://medium.freecodecamp.org/an-overview-of-the-gradient-descent-algorithm-8645c9e4de1e|An overview of the Gradient Descent algorithm]]
  * [[https://towardsdatascience.com/building-a-logistic-regression-in-python-step-by-step-becd4d56c9c8|Building A Logistic Regression in Python, Step by Step]]
  * [[https://medium.com/swlh/building-linear-regression-in-python-75a429b0d3ba|Building Linear Regression in Python]]
  * [[https://towardsdatascience.com/17-types-of-similarity-and-dissimilarity-measures-used-in-data-science-3eb914d2681|17 types of similarity and dissimilarity measures used in data science]] : explains various methods for computing distances and showing their instances in our daily lives. Additionally, it will introduce you to the pydist2 package. Mahmoud Harmouch, Medium, 14/03/2021
  * [[http://jakevdp.github.io/|Pythonic Perambulations]] de Jake VanderPlas




----

===== Synopsis (français) =====
(Méthodes de calcul appliqué à la chimie)

  * Learning outcomes teaching unit (UE):
    * Apply standard numerical methods or existing software to solve numerical problems related to scientific research activities
    * Be active in the search for existing numerical methods adapted to problems encountered by chemists
  * Content of the UE:
    * widespread methods : linear systems, numerical integration, root findings
    * Ordinary differential equations (numerical solutions, kinetic applications,...)
    * Partial differential equations (finite differences, diffusion problems)
    * Nonlinear systems of equations (Newton-Raphson method)
    * Eigenvalues eignevectors problems (applications to relaxation and population problems)
    * Approximation by linear and non-linear least squares methods (maximum likelihood, application to deconvolution)
    * Chebyshev approximation
    * Molecular modeling and visualization
    * Minimization and conformational problems
  * Prerequisite skills
    * Basic knowledge of a programming language
    * Basics of Mathematics
  * Exercises and applications: codes written in or mainly written in Python, with the general libraries matplotlib, numpy, scipy, pandas as well as other specialized libraries, especially in chemistry
  * Types of evaluations: Oral examination based on an in-depth study on one of the chapters of the course or an additional theme

  * Acquis d'apprentissage UE :
    * Appliquer des méthodes numériques standards ou des logiciels existant pour résoudre des problèmes fondamentaux ou annexes, liés à des activités de recherche scientifique
    * Être actif dans la recherche de méthodes de résolution numérique existantes et adaptées à des problèmes auxquels les chimistes sont confrontés
  * Contenu de l'UE :
    * Équations différentielles ordinaires (résolutions numériques et applications cinétiques)
    * Équations aux dérivées partielles (différences finies, problèmes de diffusion)
    * Systèmes d’équations non linéaires (méthode de Newton-Raphson)
    * Problèmes aux valeurs propres (applications à des problèmes de relaxation et de population)
    * Approximation par moindre carrés linéaires et non-linéaires (application à la déconvolution)
    * Approximation de Tchébyshev
    * Modélisation et visualisation de molécules
    * Minimisation et problèmes conformationnels
  * Compétences préalables
    * Connaissance de base d'un langage de programmation
    * Bases des mathématiques
  * Exercices et applications : codes écrits ou à écrire principalement en Python, avec les librairies générales matplotlib, numpy et scipy, ainsi que d'autres librairies spécialisées, notamment en chimie
  * Types d'évaluations : Examen oral sur base d'un travail approfondi sur un des chapitres du cours ou un thème additionnel

==== Pré-requis mathématiques ====

==== Base de la programmation ====

==== Méthodes numériques de base ====
  * [[system_of_linear_equations|Systèmes d'équations linéaires]]
    * Diagonalisation et triangularisation
    * Décomposition LU en matrices triangulaires
  * [[numerical_integration|Intégration numérique]]
    * Simpson et quadratures gaussiennes
  * [[root-finding_algorithm|Résolutions d'équations du type f(x) = 0]]
    * Équations polynomiales
    * Recherche dichotomique
    * Méthode de la sécante
    * Méthode de Newton-raphson

==== Méthodes numériques usuelles ====
=== Équations différentielles ordinaires ===

[[numerical_methods_for_ordinary_differential_equations|Résolutions numériques des ODE]]
  * principe de discrétisation, méthode d'Euler
  * Améliorations et méthodes de Runge-Kutta
  * Runge-Kutta d'ordre 4
  * Contrôle du pas d'intégration
  * Méthodes predictor-corrector
  * Méthodes d'extrapolation (Richardson, Burlish-Stoer)
  * applications :
    * équations de cinétique chimique
    * Équation logistique
      * [[wp>fr:Modèle_de_Verhulst|Modèle de Verhulst]], [[wp>fr:Dynamique_des_populations|dynamique des populations]] et non-linéarité
      * Bifurcation, doublements de périodes et transition vers le chaos
    * Réactions chimiques oscillantes : Belousov-Zhabotinsky, Brusselator, Oregonator
    * Modèle proie-prédateur
    * Attracteur étrange, modèle atmosphérique de Lorenz

=== Équations aux dérivées partielles ===
[[partial_differential_equation|Résolutions numériques des équations aux dérivées partielles]]
  * Domaine d'application des équations : équation de diffusion, équation d'ondes, équations de Navier-Stokes
  * Types de traitements numériques
  * Différences finies et problèmes de diffusion
  * Schémas classiques de différences finies
    * Résolutions stationnaires
    * Résolutions dépendantes du temps
  * Méthodes explicites, critère de (ou d'in)stabilités et méthodes implicites


=== Problèmes aux valeurs propres ===
[[eigenvalues_and_eigenvectors|Valeurs propres et vecteurs propres]]

applications à des problèmes de relaxation et de population, analyse de modes normaux de vibration, PCA (principal component analysis),...

=== Systèmes d’équations non linéaires ===
Méthode de Newton-Raphson


=== Approximation par moindre carrés linéaires et non-linéaires ===
application à la déconvolution

=== Approximations de Tchébyshev ===

=== Modélisation et visualisation de molécules ===

=== Minimisation ===
problèmes conformationnels

----

===== Thèmes additionnels =====
  * Bioinformatique et algorithmes spécifiques
  * Chimie
    * calculs quantiques, de minimisation, de mécanique moléculaire
    * représentations
  * Data science, statistiques (librairie Python Pandas,...)
    * Time series analysis
    * Machine learning (Scikit-learn,...)
  * Data visualization
    * boxplot, 3D, animations, graphes,...
  * Senseurs et interfaçage, Arduino, Raspberry Pi, IoT
  * Simulations
    * Agent base modelling et systèmes complexes
    * Automates cellulaires
    * Simpy,...
  * Traitement d'image
    * particle tracking,...