3

I know a bit of LaTeX but struggle with making large multi-line equations look professional. Apologies if this is the wrong way/place to ask. Literally any help would be grand. Specifically I would like the formatting to be on the left.

(Here is some LaTeX for reference...)

\begin{proof} First we prove \eqref{E-alpha-1}. \begin{align} \left|\mathbf{E}_{\alpha,1}^{(\beta)}( t) v\right|_{\mu} = \sum_{j=1}^{\infty}\lambda_j^{\mu}\langle(E_{\alpha,1}^{(\beta)}(t)v,e_j\rangle^2 \end{align} Note that $\langle(\mathbf{E}_{\alpha,1}^{(\beta)}t)v,e_j\rangle = \sum_{k=0}^{\infty}\frac{\langle (-A)^{\beta k}t^{\alpha k}v,e_j\rangle}{\Gamma(\alpha k+1)} = \sum_{k=0}^{\infty}\frac{\langle (-\lambda_j)^{\beta k}t^{\alpha k}v,e_j\rangle}{\Gamma(\alpha k+1)} = E_{\alpha,1}(-\lambda_j^{\beta} t^{\alpha})\langle v, e_j \rangle$. Using that $A^{\beta k}e_j = \lambda_j^{\beta k} e_j$ whenever $e_j$ are the eigenvectors of $A$. Next we use the result that $|E_{\alpha,1}(z)|^2\leq \frac{C}{(1+|z|)^2}$ Thus \begin{align} |\mathbf{E}_{\alpha,1}^{(\beta)}( t) v|_{\mu}^2 = \sum_{j=1}^{\infty}\lambda_j^{\mu}\langle(E_{\alpha,1}^{(\beta)}(t)v,e_j\rangle^2 = \sum_{j=1}^{\infty}\lambda_j^{\mu}\langle v, e_j \rangle^2 |E_{\alpha,1}(-\lambda_j^{\beta} t^{\alpha})|^2 \\ \leq \sum_{j=1}^{\infty}\lambda_j^{\mu}\langle v, e_j \rangle^2 \frac{1}{(1+\lambda_j^{\beta}t^{\alpha})^2}\\ \leq C t^{-\frac{\alpha}{\beta}(\mu - \nu)} \sum_{j=1}^{\infty}\frac{(\lambda_j^\beta t^{\alpha})^{\frac{\mu-\nu}{\beta}}}{(1+\lambda_j^{\beta}t^{\alpha})^2}\lambda^{\nu}_j\langle v, e_j \rangle^2 \\ \leq Ct^{-\frac{\alpha(\mu-\nu)}{\beta}}\sum_{j=1}^{\infty}\lambda_j^{\nu}\langle v, e_j\rangle^2 \leq Ct^{-\frac{\alpha(\mu-\nu)}{\beta}}|v|_{\nu}^2 \end{align} \end{proof} ) 
New contributor
George Dewhirst is a new contributor to this site. Take care in asking for clarification, commenting, and answering. Check out our Code of Conduct.
3
  • 5
    It would be better if you could write a minimal working example, code that would compile. Also, since you are asking about aesthetics of typesetting, a picture of your output would be nice. And you can enclose your code in triple backticks for a formatted code display. Commented Nov 24 at 22:47
  • 3
    you need to mark the code as a code block {} in the editor or use ``` before and after it, also show a complete document, so people can run your code and see the problem you are asking about Commented Nov 24 at 22:53
  • Welcome to TeX.SE! Commented Nov 25 at 0:37

2 Answers 2

5

Perhaps you had something like the code below in mind. I removed the equation numbering, but if you want to bring it back, simply replace align* with align. The fleqn option allows you to left-align equations. Also, since my crystal ball seems to have gone missing, I included in the preamble only what your formulas required.

\documentclass[12pt,fleqn]{article} % fleqn = формулы прижаты к левому краю \usepackage{amsmath,amssymb,amsthm} \newtheorem{theorem}{Theorem} \begin{document} \begin{proof} First we prove \eqref{E-alpha-1} \begin{align*} \left|\mathbf{E}_{\alpha,1}^{(\beta)}( t) v\right|_{\mu} &= \sum_{j=1}^{\infty} \lambda_j^{\mu} \langle (E_{\alpha,1}^{(\beta)}(t)v), e_j \rangle^2. \end{align*} Note that \begin{align*} \langle (\mathbf{E}_{\alpha,1}^{(\beta)}t)v, e_j \rangle &= \sum_{k=0}^{\infty} \frac{\langle (-A)^{\beta k} t^{\alpha k} v, e_j \rangle} {\Gamma(\alpha k+1)} \\[6pt] &= \sum_{k=0}^{\infty} \frac{\langle (-\lambda_j)^{\beta k} t^{\alpha k} v, e_j \rangle} {\Gamma(\alpha k+1)} \\[6pt] &= E_{\alpha,1}(-\lambda_j^{\beta} t^{\alpha}) \langle v, e_j \rangle . \end{align*} Using that $A^{\beta k} e_j = \lambda_j^{\beta k} e_j$ whenever $e_j$ are the eigenvectors of $A$. Next we use the result that \[ |E_{\alpha,1}(z)|^2 \le \frac{C}{(1+|z|)^2}. \] Thus \begin{align*} |\mathbf{E}_{\alpha,1}^{(\beta)}( t) v|_{\mu}^2 &= \sum_{j=1}^{\infty} \lambda_j^{\mu} \langle (E_{\alpha,1}^{(\beta)}(t)v), e_j \rangle^2 \\[6pt] &= \sum_{j=1}^{\infty} \lambda_j^{\mu} \langle v, e_j \rangle^2 |E_{\alpha,1}(-\lambda_j^{\beta} t^{\alpha})|^2 \\[6pt] &\le \sum_{j=1}^{\infty} \lambda_j^{\mu} \langle v, e_j \rangle^2 \frac{1}{(1+\lambda_j^{\beta} t^{\alpha})^2} \\[6pt] &\le C t^{-\frac{\alpha}{\beta}(\mu - \nu)} \sum_{j=1}^{\infty} \frac{(\lambda_j^\beta t^{\alpha})^{\frac{\mu-\nu}{\beta}}} {(1+\lambda_j^{\beta}t^{\alpha})^2} \lambda^{\nu}_j \langle v, e_j \rangle^2 \\[6pt] &\le C t^{-\frac{\alpha(\mu-\nu)}{\beta}} \sum_{j=1}^{\infty} \lambda_j^{\nu} \langle v, e_j \rangle^2 \\[6pt] &\le C t^{-\frac{\alpha(\mu-\nu)}{\beta}} |v|_{\nu}^2 . \end{align*} \end{proof} \end{document} 

enter image description here

4

Some pointers:

  1. It is not clear what you mean by "formatting to be on the left." If you mean equation numbers on the left, you can use leqno option, if you want the displayed equations not be centered but left-aligned (with some margin), you can use the fleqn option with the documentclass, as in the below code.

  2. You use the align environment but have not set any alignment points. I assumed you want alignment at leq signs right after the line breaks; hence the example below.

  3. If your proof ends with an equation, use qedhere at the end for better placement of the end-of-proof sign.

  4. Your inline equation after (1) is too long to fit on a line. I put it in an align* environment.

  5. Equations are part of sentences, so they need proper punctuation. I added a few missing ones.

\documentclass[fleqn,leqno]{article} \usepackage{amsmath} \usepackage{amsthm} \begin{document} \begin{proof} First we prove \eqref{E-alpha-1}. \begin{align} \left|\mathbf{E}_{\alpha,1}^{(\beta)}( t) v\right|_{\mu} = \sum_{j=1}^{\infty}\lambda_j^{\mu}\langle(E_{\alpha,1}^{(\beta)}(t)v,e_j\rangle^2. \end{align} Note that \begin{align*} \langle(\mathbf{E}_{\alpha,1}^{(\beta)}t)v,e_j\rangle &= \sum_{k=0}^{\infty}\frac{\langle (-A)^{\beta k}t^{\alpha k}v,e_j\rangle}{\Gamma(\alpha k+1)} \\&= \sum_{k=0}^{\infty}\frac{\langle (-\lambda_j)^{\beta k}t^{\alpha k}v,e_j\rangle}{\Gamma(\alpha k+1)} = E_{\alpha,1}(-\lambda_j^{\beta} t^{\alpha})\langle v, e_j \rangle \end{align*} using that $A^{\beta k}e_j = \lambda_j^{\beta k} e_j$ whenever $e_j$ are the eigenvectors of $A$. Next we use the result that $|E_{\alpha,1}(z)|^2\leq \frac{C}{(1+|z|)^2}$. Thus, \begin{align} |\mathbf{E}_{\alpha,1}^{(\beta)}( t) v|_{\mu}^2 &= \sum_{j=1}^{\infty}\lambda_j^{\mu}\langle(E_{\alpha,1}^{(\beta)}(t)v,e_j\rangle^2 = \sum_{j=1}^{\infty}\lambda_j^{\mu}\langle v, e_j \rangle^2 |E_{\alpha,1}(-\lambda_j^{\beta} t^{\alpha})|^2 \\ & \leq \sum_{j=1}^{\infty}\lambda_j^{\mu}\langle v, e_j \rangle^2 \frac{1}{(1+\lambda_j^{\beta}t^{\alpha})^2}\\ & \leq C t^{-\frac{\alpha}{\beta}(\mu - \nu)} \sum_{j=1}^{\infty}\frac{(\lambda_j^\beta t^{\alpha})^{\frac{\mu-\nu}{\beta}}}{(1+\lambda_j^{\beta}t^{\alpha})^2}\lambda^{\nu}_j\langle v, e_j \rangle^2 \\ & \leq Ct^{-\frac{\alpha(\mu-\nu)}{\beta}}\sum_{j=1}^{\infty}\lambda_j^{\nu}\langle v, e_j\rangle^2 \leq Ct^{-\frac{\alpha(\mu-\nu)}{\beta}}|v|_{\nu}^2. \qedhere \end{align} \end{proof} \end{document} 

enter image description here

1
  • align is meant to be used only for multi-line equations. Equation (1) is only a single line, so should be entered with the equation environment. Commented Nov 25 at 6:02

You must log in to answer this question.

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.