[Rivet-svn] r3837 - in schools/2011-YETI: . handouts handouts/day1

[blackhole at projects.hepforge.org]

Author: hoeth
Date: Tue Jul 17 15:07:47 2012
New Revision: 3837

Log:
YETI 2011

Added:
   schools/2011-YETI/
   schools/2011-YETI/handouts/
   schools/2011-YETI/handouts/day1/
   schools/2011-YETI/handouts/day1/Sh-day1.tex

Added: schools/2011-YETI/handouts/day1/Sh-day1.tex
==============================================================================
--- /dev/null	00:00:00 1970	(empty, because file is newly added)
+++ schools/2011-YETI/handouts/day1/Sh-day1.tex	Tue Jul 17 15:07:47 2012	(r3837)
@@ -0,0 +1,332 @@
+\documentclass[a4paper,10pt]{scrartcl}
+
+\usepackage{fullpage}
+\usepackage{amsmath}
+\usepackage{helvet}
+\usepackage{url}
+
+\setlength{\parindent}{0in}
+\newcommand{\done}{{\rm d}}
+\newcommand{\nnb}{\nonumber}
+
+%opening
+\title{YETI 2011 Tutorials: \\ Day 1 Sherpa -- V+Jets}
+\date{}
+
+\begin{document}
+
+\maketitle
+
+\section{Introduction}
+
+For today's session, you will be first playing a bit with Sherpa and
+then working in small groups to create data for Z+jets events. At the
+end of the tutorial you will combine your results and discuss them. 
+
+\subsection{About Sherpa}
+
+Sherpa is a full-featured event generator which puts its emphasis on an
+improved description of the perturbative stages of event generation,
+i.e. the hard scattering process described by matrix elements and the
+parton shower stage with its resummation of soft and collinear
+enhancements.
+
+One main ingredient towards that aim is the generation of hard QCD emissions
+with an exact matrix element, because the parton shower approximation is
+not valid in that case. To run a parton-shower on top of such a matrix
+element which potentially already contains hard emissions, a prescription
+called ``CKKW merging'' is implemented. Information about this merging
+can be found in the Sherpa publication at arXiv:0811.4622 and in more
+detail at arXiv:0903.1219 and its references. In addition to these
+perturbative event phases, Sherpa also has a cluster fragmentation
+module and hadron decays including QED radiation resummed in the YFS
+approach.
+
+While you work through these tutorials today, Sherpa authors will be
+available for all your questions and comments. Please don't hesitate to
+ask us.
+
+
+\subsection{Run card}
+
+The way a particular simulation runs in Sherpa is defined by several
+parameters, which can all be listed in a common file. This default name
+of the steering file is Run.dat, but other names can be used. The first
+step in running Sherpa is to adjust all parameters to the needs of the
+desired simulation.  Instructions for properly constructing these files
+are given in the Sherpa manual
+(\url{http://projects.hepforge.org/sherpa/doc/SHERPA-MC-1.2.3.html}),
+but for now we will discuss a couple of the most important features.
+
+In the ``processes'' section of the Run.dat file, the hard scattering
+processes that will be simulated are specified. The particles are
+identifed by their PDG codes. There are also so-called particle
+containers, which allow you to specify several processes with one line.
+For example, the particle container for jets, "93", includes all
+processes with
+$d,\bar{d},u,\bar{u},s,\bar{s},c,\bar{c},b,\bar{b},g$ in this place. A
+list of particle codes and particle containers is displayed when Sherpa
+is run.
+
+For all steps in these tutorials, we have prepared Run.dat files.
+
+\subsection{Getting started: V+jets @ Tevatron}
+
+Go into today's working directory:
+
+\begin{verbatim}
+  cd ~/day1/sherpa/intro
+\end{verbatim}
+
+The run card in this directory contains lots of options and comments.
+Look at the run card to find the following information about the run:
+
+\begin{itemize}
+\item Beam settings
+\item Hard scattering process:
+  \begin{itemize}
+  \item Which physics process are we running? Which lepton is being
+        produced (Hint: The PDG code for electrons is 11, while the
+        electron neutrino is 12)? 
+  \item Up to how many hard jets are being accounted for by exact matrix
+        elements (Hint: Look for curly brackets)?
+  \item Are any cuts imposed on the hard scattering (Hint: Look at the
+        ``selector'' section)? Why?
+  \end{itemize}
+\end{itemize}
+
+When you run Sherpa for the first time, it will integrate the cross
+sections.  Depending on the hard processes specified in the run card
+this may take a rather long time. The integration results can be saved
+and re-used in later runs, for this the directory {\tt Results} has to
+be created before running Sherpa. If you want to use a different
+directory, you can define it in the run card by setting the parameter
+{\tt RESULT\_DIRECTORY}.
+
+\textbf{Warning:} When you change relevant parameters in {\tt Run.dat}
+these integration results will have to be deleted by hand so that they
+get re-generated.
+
+Now it is time to run Sherpa. This is as easy as typing
+
+\begin{verbatim}
+  Sherpa
+\end{verbatim}
+
+Now you can switch to {\tt EVENTS=1} and {\tt OUTPUT=3} and actually
+look at an event printed on screen. 
+
+\textbf{Hint:} You can set all parameters also from the command line
+(overwriting the ones in Run.dat, e.g. {\tt Sherpa EVENTS=3}).
+
+\textbf{Hint:} If you specified a large number of events and don't want
+to wait to the end, you can abort with {\tt Ctrl-C} and Sherpa will
+write the output for the statistics it has generated up to that point.
+
+\subsection{Plotting a simple observable}
+
+Although Sherpa also has an internal analysis framework, we will use
+Rivet for the analyses of this tutorial. Rivet is run internally and
+writes the results to a file with extension {\tt .aida}. The default
+name is {\tt Analysis.aida}, but you can use the {\tt ANALYSIS\_OUTPUT}
+parameter in your run card to define a more useful name. In the {\tt
+(analysis)} section the analyses that Rivet will perform are listed.
+Today we will use a generic W+jets analysis. The histograms can be
+plotted using the command
+
+\begin{verbatim}
+  rivet-mkhtml -o ~/www/plots Analysis.aida
+\end{verbatim}
+
+This creates a directory called {\tt \~{}/www/plots}, stores the histograms
+there and links them in a {\tt html} page. You can then look at them by
+opening the page in your browser:
+\url{http://www.ippp.dur.ac.uk/~guestXX/} (assuming that your login name
+is {\tt guestXX} -- change this accordingly).
+
+You can plot several histograms together by listing more {\tt .aida}
+files in the above command.
+
+If the folder {\tt \~{}/www/plots} already exists because you have plotted
+results before, you have to delete it ({\tt rm -r \~{}/www/plots/}) before
+plotting again. Alternatively you can specify another output directory.
+
+
+\section{Z+jets on ME Level}
+
+\subsection{Physics}
+
+%needs more explanations
+The signal process in event generation is calculated perturbatively
+using matrix elements. In this section of the tutorial, we will look at
+the effects on observables of adding additional hard radiation in the
+matrix element to production of Z-bosons.
+
+\subsection{Running Sherpa}
+
+The setups for this section can be found in the folder {\tt \~{}/day1/sherpa/ME}.
+
+\begin{verbatim}
+  cd ~/day1/sherpa/ME
+  ls
+\end{verbatim}
+
+You will find four different run cards named {\tt Run.Xjet.dat}, where
+'X' is between 0 and 3. Take a look at the run cards. In particular,
+look at the (processes) section to see the number of additional jets
+produced by the matrix element. Also inspect the shower settings, which
+basically disable the shower.  The cross sections have already been
+integrated for you. The results are in the folders {\tt Results.Xjet}.
+These are already set in the run cards. 
+
+Once you are satisfied, run Sherpa and Rivet using the commands below. 
+\begin{verbatim}
+  Sherpa RUNDATA=Run.Xjet.dat EVENTS=100000
+\end{verbatim}
+where the `X' in {\tt Xjet} is the relevant number between 0 and 3.
+Coordinate this with the other members of your group so that not
+everybody has to run all jet multiplicities.
+
+\subsection{Plotting your Results}
+
+Collect results from other members of your group by copying all relevant
+`.aida' files to a common directory (use cp to transfer files from
+another user's home directory).
+
+To plot your results, enter the following command and look at the
+output at \url{http://www.ippp.dur.ac.uk/~guestXX/}:
+
+\begin{verbatim}
+  rivet-mkhtml -o ~/www/ME-plots Analysis.3jet.aida Analysis.2jet.aida \
+      Analysis.1jet.aida Analysis.0jet.aida
+\end{verbatim}
+
+If the folder {\tt \~{}/www/ME-plots} already exists because you have
+plotted results before, you have to delete it ({\tt rm -r \~{}/www/ME-plots})
+before plotting again.
+
+Why is the agreement with data in the low $p_\perp$ region so poor?
+Can you see where the effect of multi-jet events shows up?
+
+For the ``max0jet'' sample, one would naively expect the $p_\perp$ of
+the $Z$ boson to be $0$ (why?). But as you should see, it does get a
+very soft $p_\perp$ kick. Do you have any idea where that comes from?
+
+
+\section{Z+jets with ME/PS Merging}
+
+\subsection{Physics}
+
+Monte Carlo event generators generally rely on separating events into
+different stages. As mentioned above, the hard interaction is calculated
+perturbatively using the matrix element approach. However, the
+computational work required for this increases approximately factorially
+with the order, so it is not realistically possible to calculate
+high-multiplicity events using purely this method.
+
+The parton shower describes the soft and collinear emissions from final
+state partons by resumming the leading logarithmic terms. However, as
+the non-leading terms are neglected, the parton shower does not describe
+hard or wide-angled parton emission well. 
+
+Therefore, the multi-jet phase space is separated into two regions, with
+the hard, wide-angled emissions described by the matrix element, and the
+soft, collinear emissions described by the parton shower. Sherpa employs
+a procedure called CKKW merging to combine the matrix elements with the
+parton shower, while avoiding double-counting of phase space, and
+minimizing the dependence on the choice of phase space cut.
+
+This tutorial will give you the opportunity to compare the radiation
+patterns produced from tree-level matrix element calculations with the
+corresponding parton shower results, by comparing results of event
+generation with a maximum of 0, 1, 2, or 3 jets in the matrix element. 
+
+{\bf Please note:} {\it The comparison of results for different numbers
+of jets in the matrix element is just an exercise for this tutorial.
+When using ME/PS merging it is always advisable to have as many jets as
+computationally possible in the matrix element.}
+
+
+\subsection{Running Sherpa}
+
+Within your group, decide which jet multiplicities each member will run.
+Remember that the higher the jet multiplicity, the longer the run will take.
+
+The setups can be found in the directory named {\tt \~{}/day1/sherpa/merging}.
+
+Take a look at the run cards. In particular, look at the (processes)
+section, and check that you are generating events with the correct
+number of jets in the final state.
+
+Once you are satisfied, run Sherpa and Rivet using the command below.
+
+\begin{verbatim}
+  Sherpa RUNDATA=Run.Xjet.dat EVENTS=100000
+\end{verbatim}
+
+\noindent where the `X' in {\tt Xjet} is the relevant number between 0 and 3.
+
+\subsection{Plotting your Results}
+
+Collect results from other members of your group by copying all relevant
+`.aida' files to a common directory and plot them like you did in the
+previous section:
+
+\begin{verbatim}
+  rivet-mkhtml -o ~/www/merging-plots Analysis.3jet.aida Analysis.2jet.aida \
+      Analysis.1jet.aida Analysis.0jet.aida
+\end{verbatim}
+
+What has changed by adding the shower? Can you explain the remaining
+differences between the runs?
+
+\section{QED Radiation}
+
+\subsection{Physics}
+
+As well as the QCD effects that produce jets, there are also QED effects
+from radiated photons. In this part of the tutorial, we are going to
+look at the effect of this QED radiation. In the YFS formalism used by
+Sherpa, the external lepton lines are dressed with resummed collinear
+photon radiation. The hardest emission is corrected to the exact matrix
+element, but the cross section is not affected.
+
+\subsection{Running Sherpa}
+
+The two setups can be found in the folder {\tt \~{}/day1/sherpa/QED}.
+Within your group, decide which setup each member will run. Since we are
+not asking for extra jets this time, your job has to integrate the cross
+section.
+
+Take a look at the run cards. In particular, check in the {\tt (me)}
+section to see if QED radiation is included.
+
+Run Sherpa and Rivet using the command:
+
+\begin{verbatim}
+  Sherpa RUNDATA=Run.*.dat EVENTS=100000
+\end{verbatim}
+
+where the * is either `ME' or `Off'. Here we are running two
+pseudo-analyses (i.e. without reference data), on the one hand to
+compare the runs, and on the other hand to compare two different
+analyses: In {\tt MC\_ZJETS}, the QED radiation in a cone around the
+lepton has been accounted for in the $Z$ reconstruction, while in {\tt
+MC\_ZJETS\_NOCLUS} this radiation has been ignored. What difference do
+you expect?
+
+\subsection{Plotting your Results}
+
+To show the results of {\tt MC\_ZJETS} and {\tt MC\_ZJETS\_NOCLUS} in
+the same plot we need to mangle the output files a bit before we
+generate the plots. The {\tt mangle\_output} script does this for you:
+
+\begin{verbatim}
+  ./mangle_output
+  rivet-mkhtml -o ~/www/QED-plots ME.aida ME_noclus.aida Off.aida Off_noclus.aida
+\end{verbatim}
+
+Which differences can you see between the runs and analyses in the
+different observables?
+
+\end{document}