3
0
mirror of https://github.com/triqs/dft_tools synced 2024-11-01 19:53:45 +01:00
dft_tools/doc/reference/c++/learn/first_mc.rst
Olivier Parcollet f2c7d449cc First commit : triqs libs version 1.0 alpha1
for earlier commits, see TRIQS0.x repository.
2013-07-17 19:24:07 +02:00

65 lines
2.0 KiB
ReStructuredText

A first external code
=====================
.. highlight:: c
As a first exercise you can try to write a Monte Carlo code for an Ising chain
in a magnetic field. Your goal is to write this code as an external project and
to use the Monte Carlo class provided by TRIQS.
Take some time to read the :ref:`Monte Carlo <montecarlo>` chapter, but don't
read the complete example at the end of the chapter because it is precisely
what you need to do here. You can check your implementation later.
.. _isingex:
Ising chain in magnetic field
-----------------------------
Here's the Hamiltonian for the problem of Ising spins in a magnetic field
.. math::
\mathcal{H} = -J \sum_{i=1}^N \sigma_i \sigma_{i+1} - h \sum_{i=1}^N \sigma_i.
The goal is to find the magnetization per spin :math:`m` of the system for
:math:`J = -1.0`, a magnetic field :math:`h = 0.5` as a function of
the inverse temperature :math:`\beta`. You can see how the results
change with the length of the chain :math:`N`.
Implementation hints
--------------------
Here are a couple of implementation hints that you might want to follow.
* In most Monte Carlo programs there is a *configuration* which is modified
along the simulation. Take enough time to think how this configuration
can be efficiently described and implement it in a separate file, say
:file:`configuration.hpp`. In this example, the configuration is a
collection of spins that can e.g. be described by a vector of integers.
+1 would be a spin up and -1 a spin down. If you're worried with memory
space, you could use a vector of booleans (true for up spins, false for
down spins).
* More to come...
Solution
--------
In the limit :math:`N \rightarrow \infty`, the solution for the magnetization
is
.. math::
m = \frac{\sinh(\beta h) + \frac{\sinh(\beta h)\cosh(\beta h)}{\sqrt{\sinh^2(\beta h) + e^{-4\beta J}}}}
{\cosh(\beta h) + \sqrt{\sinh^2(\beta h) + e^{-4\beta J}}}.
Here's a plot of :math:`m` versus :math:`\beta` for different values of :math:`N`:
.. image:: m_vs_beta.png
:width: 700
:align: center