ViewVC Help
View File | Revision Log | Show Annotations | View Changeset | Root Listing
root/group/trunk/tengDissertation/Introduction.tex
Revision: 2703
Committed: Wed Apr 12 04:13:16 2006 UTC (18 years, 2 months ago) by tim
Content type: application/x-tex
File size: 45377 byte(s)
Log Message:
minor fix

File Contents

# User Rev Content
1 tim 2685 \chapter{\label{chapt:introduction}INTRODUCTION AND THEORETICAL BACKGROUND}
2    
3 tim 2693 \section{\label{introSection:classicalMechanics}Classical
4     Mechanics}
5 tim 2685
6 tim 2692 Closely related to Classical Mechanics, Molecular Dynamics
7     simulations are carried out by integrating the equations of motion
8     for a given system of particles. There are three fundamental ideas
9     behind classical mechanics. Firstly, One can determine the state of
10     a mechanical system at any time of interest; Secondly, all the
11     mechanical properties of the system at that time can be determined
12     by combining the knowledge of the properties of the system with the
13     specification of this state; Finally, the specification of the state
14     when further combine with the laws of mechanics will also be
15     sufficient to predict the future behavior of the system.
16 tim 2685
17 tim 2693 \subsection{\label{introSection:newtonian}Newtonian Mechanics}
18 tim 2694 The discovery of Newton's three laws of mechanics which govern the
19     motion of particles is the foundation of the classical mechanics.
20     Newton¡¯s first law defines a class of inertial frames. Inertial
21     frames are reference frames where a particle not interacting with
22     other bodies will move with constant speed in the same direction.
23     With respect to inertial frames Newton¡¯s second law has the form
24     \begin{equation}
25     F = \frac {dp}{dt} = \frac {mv}{dt}
26     \label{introEquation:newtonSecondLaw}
27     \end{equation}
28     A point mass interacting with other bodies moves with the
29     acceleration along the direction of the force acting on it. Let
30 tim 2702 $F_{ij}$ be the force that particle $i$ exerts on particle $j$, and
31     $F_{ji}$ be the force that particle $j$ exerts on particle $i$.
32 tim 2694 Newton¡¯s third law states that
33     \begin{equation}
34 tim 2702 F_{ij} = -F_{ji}
35 tim 2694 \label{introEquation:newtonThirdLaw}
36     \end{equation}
37 tim 2692
38 tim 2694 Conservation laws of Newtonian Mechanics play very important roles
39     in solving mechanics problems. The linear momentum of a particle is
40     conserved if it is free or it experiences no force. The second
41     conservation theorem concerns the angular momentum of a particle.
42     The angular momentum $L$ of a particle with respect to an origin
43     from which $r$ is measured is defined to be
44     \begin{equation}
45     L \equiv r \times p \label{introEquation:angularMomentumDefinition}
46     \end{equation}
47     The torque $\tau$ with respect to the same origin is defined to be
48     \begin{equation}
49     N \equiv r \times F \label{introEquation:torqueDefinition}
50     \end{equation}
51     Differentiating Eq.~\ref{introEquation:angularMomentumDefinition},
52     \[
53     \dot L = \frac{d}{{dt}}(r \times p) = (\dot r \times p) + (r \times
54     \dot p)
55     \]
56     since
57     \[
58     \dot r \times p = \dot r \times mv = m\dot r \times \dot r \equiv 0
59     \]
60     thus,
61     \begin{equation}
62     \dot L = r \times \dot p = N
63     \end{equation}
64     If there are no external torques acting on a body, the angular
65     momentum of it is conserved. The last conservation theorem state
66 tim 2696 that if all forces are conservative, Energy
67     \begin{equation}E = T + V \label{introEquation:energyConservation}
68     \end{equation}
69     is conserved. All of these conserved quantities are
70     important factors to determine the quality of numerical integration
71     scheme for rigid body \cite{Dullweber1997}.
72 tim 2694
73 tim 2693 \subsection{\label{introSection:lagrangian}Lagrangian Mechanics}
74 tim 2692
75     Newtonian Mechanics suffers from two important limitations: it
76     describes their motion in special cartesian coordinate systems.
77     Another limitation of Newtonian mechanics becomes obvious when we
78     try to describe systems with large numbers of particles. It becomes
79     very difficult to predict the properties of the system by carrying
80     out calculations involving the each individual interaction between
81     all the particles, even if we know all of the details of the
82     interaction. In order to overcome some of the practical difficulties
83     which arise in attempts to apply Newton's equation to complex
84     system, alternative procedures may be developed.
85    
86 tim 2694 \subsubsection{\label{introSection:halmiltonPrinciple}Hamilton's
87 tim 2692 Principle}
88    
89     Hamilton introduced the dynamical principle upon which it is
90     possible to base all of mechanics and, indeed, most of classical
91     physics. Hamilton's Principle may be stated as follow,
92    
93     The actual trajectory, along which a dynamical system may move from
94     one point to another within a specified time, is derived by finding
95     the path which minimizes the time integral of the difference between
96 tim 2694 the kinetic, $K$, and potential energies, $U$ \cite{tolman79}.
97 tim 2692 \begin{equation}
98     \delta \int_{t_1 }^{t_2 } {(K - U)dt = 0} ,
99 tim 2693 \label{introEquation:halmitonianPrinciple1}
100 tim 2692 \end{equation}
101    
102     For simple mechanical systems, where the forces acting on the
103     different part are derivable from a potential and the velocities are
104     small compared with that of light, the Lagrangian function $L$ can
105     be define as the difference between the kinetic energy of the system
106     and its potential energy,
107     \begin{equation}
108     L \equiv K - U = L(q_i ,\dot q_i ) ,
109     \label{introEquation:lagrangianDef}
110     \end{equation}
111     then Eq.~\ref{introEquation:halmitonianPrinciple1} becomes
112     \begin{equation}
113 tim 2693 \delta \int_{t_1 }^{t_2 } {L dt = 0} ,
114     \label{introEquation:halmitonianPrinciple2}
115 tim 2692 \end{equation}
116    
117 tim 2694 \subsubsection{\label{introSection:equationOfMotionLagrangian}The
118 tim 2692 Equations of Motion in Lagrangian Mechanics}
119    
120 tim 2700 For a holonomic system of $f$ degrees of freedom, the equations of
121 tim 2692 motion in the Lagrangian form is
122     \begin{equation}
123     \frac{d}{{dt}}\frac{{\partial L}}{{\partial \dot q_i }} -
124     \frac{{\partial L}}{{\partial q_i }} = 0,{\rm{ }}i = 1, \ldots,f
125 tim 2693 \label{introEquation:eqMotionLagrangian}
126 tim 2692 \end{equation}
127     where $q_{i}$ is generalized coordinate and $\dot{q_{i}}$ is
128     generalized velocity.
129    
130 tim 2693 \subsection{\label{introSection:hamiltonian}Hamiltonian Mechanics}
131 tim 2692
132     Arising from Lagrangian Mechanics, Hamiltonian Mechanics was
133     introduced by William Rowan Hamilton in 1833 as a re-formulation of
134     classical mechanics. If the potential energy of a system is
135     independent of generalized velocities, the generalized momenta can
136     be defined as
137     \begin{equation}
138     p_i = \frac{\partial L}{\partial \dot q_i}
139     \label{introEquation:generalizedMomenta}
140     \end{equation}
141 tim 2693 The Lagrange equations of motion are then expressed by
142 tim 2692 \begin{equation}
143 tim 2693 p_i = \frac{{\partial L}}{{\partial q_i }}
144     \label{introEquation:generalizedMomentaDot}
145     \end{equation}
146    
147     With the help of the generalized momenta, we may now define a new
148     quantity $H$ by the equation
149     \begin{equation}
150     H = \sum\limits_k {p_k \dot q_k } - L ,
151 tim 2692 \label{introEquation:hamiltonianDefByLagrangian}
152     \end{equation}
153     where $ \dot q_1 \ldots \dot q_f $ are generalized velocities and
154     $L$ is the Lagrangian function for the system.
155    
156 tim 2693 Differentiating Eq.~\ref{introEquation:hamiltonianDefByLagrangian},
157     one can obtain
158     \begin{equation}
159     dH = \sum\limits_k {\left( {p_k d\dot q_k + \dot q_k dp_k -
160     \frac{{\partial L}}{{\partial q_k }}dq_k - \frac{{\partial
161     L}}{{\partial \dot q_k }}d\dot q_k } \right)} - \frac{{\partial
162     L}}{{\partial t}}dt \label{introEquation:diffHamiltonian1}
163     \end{equation}
164     Making use of Eq.~\ref{introEquation:generalizedMomenta}, the
165     second and fourth terms in the parentheses cancel. Therefore,
166     Eq.~\ref{introEquation:diffHamiltonian1} can be rewritten as
167     \begin{equation}
168     dH = \sum\limits_k {\left( {\dot q_k dp_k - \dot p_k dq_k }
169     \right)} - \frac{{\partial L}}{{\partial t}}dt
170     \label{introEquation:diffHamiltonian2}
171     \end{equation}
172     By identifying the coefficients of $dq_k$, $dp_k$ and dt, we can
173     find
174     \begin{equation}
175     \frac{{\partial H}}{{\partial p_k }} = q_k
176     \label{introEquation:motionHamiltonianCoordinate}
177     \end{equation}
178     \begin{equation}
179     \frac{{\partial H}}{{\partial q_k }} = - p_k
180     \label{introEquation:motionHamiltonianMomentum}
181     \end{equation}
182     and
183     \begin{equation}
184     \frac{{\partial H}}{{\partial t}} = - \frac{{\partial L}}{{\partial
185     t}}
186     \label{introEquation:motionHamiltonianTime}
187     \end{equation}
188    
189     Eq.~\ref{introEquation:motionHamiltonianCoordinate} and
190     Eq.~\ref{introEquation:motionHamiltonianMomentum} are Hamilton's
191     equation of motion. Due to their symmetrical formula, they are also
192 tim 2694 known as the canonical equations of motions \cite{Goldstein01}.
193 tim 2693
194 tim 2692 An important difference between Lagrangian approach and the
195     Hamiltonian approach is that the Lagrangian is considered to be a
196     function of the generalized velocities $\dot q_i$ and the
197     generalized coordinates $q_i$, while the Hamiltonian is considered
198     to be a function of the generalized momenta $p_i$ and the conjugate
199     generalized coordinate $q_i$. Hamiltonian Mechanics is more
200     appropriate for application to statistical mechanics and quantum
201     mechanics, since it treats the coordinate and its time derivative as
202     independent variables and it only works with 1st-order differential
203 tim 2694 equations\cite{Marion90}.
204 tim 2692
205 tim 2696 In Newtonian Mechanics, a system described by conservative forces
206     conserves the total energy \ref{introEquation:energyConservation}.
207     It follows that Hamilton's equations of motion conserve the total
208     Hamiltonian.
209     \begin{equation}
210     \frac{{dH}}{{dt}} = \sum\limits_i {\left( {\frac{{\partial
211     H}}{{\partial q_i }}\dot q_i + \frac{{\partial H}}{{\partial p_i
212     }}\dot p_i } \right)} = \sum\limits_i {\left( {\frac{{\partial
213     H}}{{\partial q_i }}\frac{{\partial H}}{{\partial p_i }} -
214     \frac{{\partial H}}{{\partial p_i }}\frac{{\partial H}}{{\partial
215 tim 2698 q_i }}} \right) = 0} \label{introEquation:conserveHalmitonian}
216 tim 2696 \end{equation}
217    
218 tim 2693 \section{\label{introSection:statisticalMechanics}Statistical
219     Mechanics}
220 tim 2692
221 tim 2694 The thermodynamic behaviors and properties of Molecular Dynamics
222 tim 2692 simulation are governed by the principle of Statistical Mechanics.
223     The following section will give a brief introduction to some of the
224 tim 2700 Statistical Mechanics concepts and theorem presented in this
225     dissertation.
226 tim 2692
227 tim 2700 \subsection{\label{introSection:ensemble}Phase Space and Ensemble}
228 tim 2692
229 tim 2700 Mathematically, phase space is the space which represents all
230     possible states. Each possible state of the system corresponds to
231     one unique point in the phase space. For mechanical systems, the
232     phase space usually consists of all possible values of position and
233     momentum variables. Consider a dynamic system in a cartesian space,
234     where each of the $6f$ coordinates and momenta is assigned to one of
235     $6f$ mutually orthogonal axes, the phase space of this system is a
236     $6f$ dimensional space. A point, $x = (q_1 , \ldots ,q_f ,p_1 ,
237     \ldots ,p_f )$, with a unique set of values of $6f$ coordinates and
238     momenta is a phase space vector.
239    
240     A microscopic state or microstate of a classical system is
241     specification of the complete phase space vector of a system at any
242     instant in time. An ensemble is defined as a collection of systems
243     sharing one or more macroscopic characteristics but each being in a
244     unique microstate. The complete ensemble is specified by giving all
245     systems or microstates consistent with the common macroscopic
246     characteristics of the ensemble. Although the state of each
247     individual system in the ensemble could be precisely described at
248     any instance in time by a suitable phase space vector, when using
249     ensembles for statistical purposes, there is no need to maintain
250     distinctions between individual systems, since the numbers of
251     systems at any time in the different states which correspond to
252     different regions of the phase space are more interesting. Moreover,
253     in the point of view of statistical mechanics, one would prefer to
254     use ensembles containing a large enough population of separate
255     members so that the numbers of systems in such different states can
256     be regarded as changing continuously as we traverse different
257     regions of the phase space. The condition of an ensemble at any time
258     can be regarded as appropriately specified by the density $\rho$
259     with which representative points are distributed over the phase
260     space. The density of distribution for an ensemble with $f$ degrees
261     of freedom is defined as,
262     \begin{equation}
263     \rho = \rho (q_1 , \ldots ,q_f ,p_1 , \ldots ,p_f ,t).
264     \label{introEquation:densityDistribution}
265     \end{equation}
266     Governed by the principles of mechanics, the phase points change
267     their value which would change the density at any time at phase
268     space. Hence, the density of distribution is also to be taken as a
269     function of the time.
270    
271     The number of systems $\delta N$ at time $t$ can be determined by,
272     \begin{equation}
273     \delta N = \rho (q,p,t)dq_1 \ldots dq_f dp_1 \ldots dp_f.
274     \label{introEquation:deltaN}
275     \end{equation}
276     Assuming a large enough population of systems are exploited, we can
277     sufficiently approximate $\delta N$ without introducing
278     discontinuity when we go from one region in the phase space to
279     another. By integrating over the whole phase space,
280     \begin{equation}
281     N = \int { \ldots \int {\rho (q,p,t)dq_1 } ...dq_f dp_1 } ...dp_f
282     \label{introEquation:totalNumberSystem}
283     \end{equation}
284     gives us an expression for the total number of the systems. Hence,
285     the probability per unit in the phase space can be obtained by,
286     \begin{equation}
287     \frac{{\rho (q,p,t)}}{N} = \frac{{\rho (q,p,t)}}{{\int { \ldots \int
288     {\rho (q,p,t)dq_1 } ...dq_f dp_1 } ...dp_f }}.
289     \label{introEquation:unitProbability}
290     \end{equation}
291     With the help of Equation(\ref{introEquation:unitProbability}) and
292     the knowledge of the system, it is possible to calculate the average
293     value of any desired quantity which depends on the coordinates and
294     momenta of the system. Even when the dynamics of the real system is
295     complex, or stochastic, or even discontinuous, the average
296     properties of the ensemble of possibilities as a whole may still
297     remain well defined. For a classical system in thermal equilibrium
298     with its environment, the ensemble average of a mechanical quantity,
299     $\langle A(q , p) \rangle_t$, takes the form of an integral over the
300     phase space of the system,
301     \begin{equation}
302     \langle A(q , p) \rangle_t = \frac{{\int { \ldots \int {A(q,p)\rho
303     (q,p,t)dq_1 } ...dq_f dp_1 } ...dp_f }}{{\int { \ldots \int {\rho
304     (q,p,t)dq_1 } ...dq_f dp_1 } ...dp_f }}
305     \label{introEquation:ensembelAverage}
306     \end{equation}
307    
308     There are several different types of ensembles with different
309     statistical characteristics. As a function of macroscopic
310     parameters, such as temperature \textit{etc}, partition function can
311     be used to describe the statistical properties of a system in
312     thermodynamic equilibrium.
313    
314     As an ensemble of systems, each of which is known to be thermally
315     isolated and conserve energy, Microcanonical ensemble(NVE) has a
316     partition function like,
317     \begin{equation}
318     \Omega (N,V,E) = e^{\beta TS}
319     \label{introEqaution:NVEPartition}.
320     \end{equation}
321     A canonical ensemble(NVT)is an ensemble of systems, each of which
322     can share its energy with a large heat reservoir. The distribution
323     of the total energy amongst the possible dynamical states is given
324     by the partition function,
325     \begin{equation}
326     \Omega (N,V,T) = e^{ - \beta A}
327     \label{introEquation:NVTPartition}
328     \end{equation}
329     Here, $A$ is the Helmholtz free energy which is defined as $ A = U -
330     TS$. Since most experiment are carried out under constant pressure
331     condition, isothermal-isobaric ensemble(NPT) play a very important
332     role in molecular simulation. The isothermal-isobaric ensemble allow
333     the system to exchange energy with a heat bath of temperature $T$
334     and to change the volume as well. Its partition function is given as
335     \begin{equation}
336     \Delta (N,P,T) = - e^{\beta G}.
337     \label{introEquation:NPTPartition}
338     \end{equation}
339     Here, $G = U - TS + PV$, and $G$ is called Gibbs free energy.
340    
341     \subsection{\label{introSection:liouville}Liouville's theorem}
342    
343     The Liouville's theorem is the foundation on which statistical
344     mechanics rests. It describes the time evolution of phase space
345     distribution function. In order to calculate the rate of change of
346     $\rho$, we begin from Equation(\ref{introEquation:deltaN}). If we
347     consider the two faces perpendicular to the $q_1$ axis, which are
348     located at $q_1$ and $q_1 + \delta q_1$, the number of phase points
349     leaving the opposite face is given by the expression,
350     \begin{equation}
351     \left( {\rho + \frac{{\partial \rho }}{{\partial q_1 }}\delta q_1 }
352     \right)\left( {\dot q_1 + \frac{{\partial \dot q_1 }}{{\partial q_1
353     }}\delta q_1 } \right)\delta q_2 \ldots \delta q_f \delta p_1
354     \ldots \delta p_f .
355     \end{equation}
356     Summing all over the phase space, we obtain
357     \begin{equation}
358     \frac{{d(\delta N)}}{{dt}} = - \sum\limits_{i = 1}^f {\left[ {\rho
359     \left( {\frac{{\partial \dot q_i }}{{\partial q_i }} +
360     \frac{{\partial \dot p_i }}{{\partial p_i }}} \right) + \left(
361     {\frac{{\partial \rho }}{{\partial q_i }}\dot q_i + \frac{{\partial
362     \rho }}{{\partial p_i }}\dot p_i } \right)} \right]} \delta q_1
363     \ldots \delta q_f \delta p_1 \ldots \delta p_f .
364     \end{equation}
365     Differentiating the equations of motion in Hamiltonian formalism
366     (\ref{introEquation:motionHamiltonianCoordinate},
367     \ref{introEquation:motionHamiltonianMomentum}), we can show,
368     \begin{equation}
369     \sum\limits_i {\left( {\frac{{\partial \dot q_i }}{{\partial q_i }}
370     + \frac{{\partial \dot p_i }}{{\partial p_i }}} \right)} = 0 ,
371     \end{equation}
372     which cancels the first terms of the right hand side. Furthermore,
373     divining $ \delta q_1 \ldots \delta q_f \delta p_1 \ldots \delta
374     p_f $ in both sides, we can write out Liouville's theorem in a
375     simple form,
376     \begin{equation}
377     \frac{{\partial \rho }}{{\partial t}} + \sum\limits_{i = 1}^f
378     {\left( {\frac{{\partial \rho }}{{\partial q_i }}\dot q_i +
379     \frac{{\partial \rho }}{{\partial p_i }}\dot p_i } \right)} = 0 .
380     \label{introEquation:liouvilleTheorem}
381     \end{equation}
382    
383     Liouville's theorem states that the distribution function is
384     constant along any trajectory in phase space. In classical
385     statistical mechanics, since the number of particles in the system
386     is huge, we may be able to believe the system is stationary,
387     \begin{equation}
388     \frac{{\partial \rho }}{{\partial t}} = 0.
389     \label{introEquation:stationary}
390     \end{equation}
391     In such stationary system, the density of distribution $\rho$ can be
392     connected to the Hamiltonian $H$ through Maxwell-Boltzmann
393     distribution,
394     \begin{equation}
395     \rho \propto e^{ - \beta H}
396     \label{introEquation:densityAndHamiltonian}
397     \end{equation}
398    
399 tim 2702 \subsubsection{\label{introSection:phaseSpaceConservation}Conservation of Phase Space}
400     Lets consider a region in the phase space,
401     \begin{equation}
402     \delta v = \int { \ldots \int {dq_1 } ...dq_f dp_1 } ..dp_f .
403     \end{equation}
404     If this region is small enough, the density $\rho$ can be regarded
405     as uniform over the whole phase space. Thus, the number of phase
406     points inside this region is given by,
407     \begin{equation}
408     \delta N = \rho \delta v = \rho \int { \ldots \int {dq_1 } ...dq_f
409     dp_1 } ..dp_f.
410     \end{equation}
411    
412     \begin{equation}
413     \frac{{d(\delta N)}}{{dt}} = \frac{{d\rho }}{{dt}}\delta v + \rho
414     \frac{d}{{dt}}(\delta v) = 0.
415     \end{equation}
416     With the help of stationary assumption
417     (\ref{introEquation:stationary}), we obtain the principle of the
418     \emph{conservation of extension in phase space},
419     \begin{equation}
420     \frac{d}{{dt}}(\delta v) = \frac{d}{{dt}}\int { \ldots \int {dq_1 }
421     ...dq_f dp_1 } ..dp_f = 0.
422     \label{introEquation:volumePreserving}
423     \end{equation}
424    
425     \subsubsection{\label{introSection:liouvilleInOtherForms}Liouville's Theorem in Other Forms}
426    
427 tim 2700 Liouville's theorem can be expresses in a variety of different forms
428     which are convenient within different contexts. For any two function
429     $F$ and $G$ of the coordinates and momenta of a system, the Poisson
430     bracket ${F, G}$ is defined as
431     \begin{equation}
432     \left\{ {F,G} \right\} = \sum\limits_i {\left( {\frac{{\partial
433     F}}{{\partial q_i }}\frac{{\partial G}}{{\partial p_i }} -
434     \frac{{\partial F}}{{\partial p_i }}\frac{{\partial G}}{{\partial
435     q_i }}} \right)}.
436     \label{introEquation:poissonBracket}
437     \end{equation}
438     Substituting equations of motion in Hamiltonian formalism(
439     \ref{introEquation:motionHamiltonianCoordinate} ,
440     \ref{introEquation:motionHamiltonianMomentum} ) into
441     (\ref{introEquation:liouvilleTheorem}), we can rewrite Liouville's
442     theorem using Poisson bracket notion,
443     \begin{equation}
444     \left( {\frac{{\partial \rho }}{{\partial t}}} \right) = - \left\{
445     {\rho ,H} \right\}.
446     \label{introEquation:liouvilleTheromInPoissin}
447     \end{equation}
448     Moreover, the Liouville operator is defined as
449     \begin{equation}
450     iL = \sum\limits_{i = 1}^f {\left( {\frac{{\partial H}}{{\partial
451     p_i }}\frac{\partial }{{\partial q_i }} - \frac{{\partial
452     H}}{{\partial q_i }}\frac{\partial }{{\partial p_i }}} \right)}
453     \label{introEquation:liouvilleOperator}
454     \end{equation}
455     In terms of Liouville operator, Liouville's equation can also be
456     expressed as
457     \begin{equation}
458     \left( {\frac{{\partial \rho }}{{\partial t}}} \right) = - iL\rho
459     \label{introEquation:liouvilleTheoremInOperator}
460     \end{equation}
461    
462 tim 2693 \subsection{\label{introSection:ergodic}The Ergodic Hypothesis}
463 tim 2692
464 tim 2695 Various thermodynamic properties can be calculated from Molecular
465     Dynamics simulation. By comparing experimental values with the
466     calculated properties, one can determine the accuracy of the
467     simulation and the quality of the underlying model. However, both of
468     experiment and computer simulation are usually performed during a
469     certain time interval and the measurements are averaged over a
470     period of them which is different from the average behavior of
471     many-body system in Statistical Mechanics. Fortunately, Ergodic
472     Hypothesis is proposed to make a connection between time average and
473     ensemble average. It states that time average and average over the
474     statistical ensemble are identical \cite{Frenkel1996, leach01:mm}.
475     \begin{equation}
476 tim 2700 \langle A(q , p) \rangle_t = \mathop {\lim }\limits_{t \to \infty }
477     \frac{1}{t}\int\limits_0^t {A(q(t),p(t))dt = \int\limits_\Gamma
478     {A(q(t),p(t))} } \rho (q(t), p(t)) dqdp
479 tim 2695 \end{equation}
480 tim 2700 where $\langle A(q , p) \rangle_t$ is an equilibrium value of a
481     physical quantity and $\rho (p(t), q(t))$ is the equilibrium
482     distribution function. If an observation is averaged over a
483     sufficiently long time (longer than relaxation time), all accessible
484     microstates in phase space are assumed to be equally probed, giving
485     a properly weighted statistical average. This allows the researcher
486     freedom of choice when deciding how best to measure a given
487     observable. In case an ensemble averaged approach sounds most
488     reasonable, the Monte Carlo techniques\cite{metropolis:1949} can be
489     utilized. Or if the system lends itself to a time averaging
490     approach, the Molecular Dynamics techniques in
491     Sec.~\ref{introSection:molecularDynamics} will be the best
492     choice\cite{Frenkel1996}.
493 tim 2694
494 tim 2697 \section{\label{introSection:geometricIntegratos}Geometric Integrators}
495     A variety of numerical integrators were proposed to simulate the
496     motions. They usually begin with an initial conditionals and move
497     the objects in the direction governed by the differential equations.
498     However, most of them ignore the hidden physical law contained
499     within the equations. Since 1990, geometric integrators, which
500     preserve various phase-flow invariants such as symplectic structure,
501     volume and time reversal symmetry, are developed to address this
502     issue. The velocity verlet method, which happens to be a simple
503     example of symplectic integrator, continues to gain its popularity
504     in molecular dynamics community. This fact can be partly explained
505     by its geometric nature.
506    
507     \subsection{\label{introSection:symplecticManifold}Symplectic Manifold}
508     A \emph{manifold} is an abstract mathematical space. It locally
509     looks like Euclidean space, but when viewed globally, it may have
510     more complicate structure. A good example of manifold is the surface
511     of Earth. It seems to be flat locally, but it is round if viewed as
512     a whole. A \emph{differentiable manifold} (also known as
513     \emph{smooth manifold}) is a manifold with an open cover in which
514     the covering neighborhoods are all smoothly isomorphic to one
515     another. In other words,it is possible to apply calculus on
516     \emph{differentiable manifold}. A \emph{symplectic manifold} is
517     defined as a pair $(M, \omega)$ which consisting of a
518     \emph{differentiable manifold} $M$ and a close, non-degenerated,
519     bilinear symplectic form, $\omega$. A symplectic form on a vector
520     space $V$ is a function $\omega(x, y)$ which satisfies
521     $\omega(\lambda_1x_1+\lambda_2x_2, y) = \lambda_1\omega(x_1, y)+
522     \lambda_2\omega(x_2, y)$, $\omega(x, y) = - \omega(y, x)$ and
523     $\omega(x, x) = 0$. Cross product operation in vector field is an
524     example of symplectic form.
525    
526     One of the motivations to study \emph{symplectic manifold} in
527     Hamiltonian Mechanics is that a symplectic manifold can represent
528     all possible configurations of the system and the phase space of the
529     system can be described by it's cotangent bundle. Every symplectic
530     manifold is even dimensional. For instance, in Hamilton equations,
531     coordinate and momentum always appear in pairs.
532    
533     Let $(M,\omega)$ and $(N, \eta)$ be symplectic manifolds. A map
534     \[
535     f : M \rightarrow N
536     \]
537     is a \emph{symplectomorphism} if it is a \emph{diffeomorphims} and
538     the \emph{pullback} of $\eta$ under f is equal to $\omega$.
539     Canonical transformation is an example of symplectomorphism in
540 tim 2698 classical mechanics.
541 tim 2697
542 tim 2698 \subsection{\label{introSection:ODE}Ordinary Differential Equations}
543 tim 2697
544 tim 2698 For a ordinary differential system defined as
545     \begin{equation}
546     \dot x = f(x)
547     \end{equation}
548     where $x = x(q,p)^T$, this system is canonical Hamiltonian, if
549     \begin{equation}
550 tim 2699 f(r) = J\nabla _x H(r).
551 tim 2698 \end{equation}
552     $H = H (q, p)$ is Hamiltonian function and $J$ is the skew-symmetric
553     matrix
554     \begin{equation}
555     J = \left( {\begin{array}{*{20}c}
556     0 & I \\
557     { - I} & 0 \\
558     \end{array}} \right)
559     \label{introEquation:canonicalMatrix}
560     \end{equation}
561     where $I$ is an identity matrix. Using this notation, Hamiltonian
562     system can be rewritten as,
563     \begin{equation}
564     \frac{d}{{dt}}x = J\nabla _x H(x)
565     \label{introEquation:compactHamiltonian}
566     \end{equation}In this case, $f$ is
567     called a \emph{Hamiltonian vector field}.
568 tim 2697
569 tim 2698 Another generalization of Hamiltonian dynamics is Poisson Dynamics,
570     \begin{equation}
571     \dot x = J(x)\nabla _x H \label{introEquation:poissonHamiltonian}
572     \end{equation}
573     The most obvious change being that matrix $J$ now depends on $x$.
574     The free rigid body is an example of Poisson system (actually a
575     Lie-Poisson system) with Hamiltonian function of angular kinetic
576     energy.
577     \begin{equation}
578     J(\pi ) = \left( {\begin{array}{*{20}c}
579     0 & {\pi _3 } & { - \pi _2 } \\
580     { - \pi _3 } & 0 & {\pi _1 } \\
581     {\pi _2 } & { - \pi _1 } & 0 \\
582     \end{array}} \right)
583     \end{equation}
584    
585     \begin{equation}
586     H = \frac{1}{2}\left( {\frac{{\pi _1^2 }}{{I_1 }} + \frac{{\pi _2^2
587     }}{{I_2 }} + \frac{{\pi _3^2 }}{{I_3 }}} \right)
588     \end{equation}
589    
590 tim 2702 \subsection{\label{introSection:exactFlow}Exact Flow}
591    
592 tim 2698 Let $x(t)$ be the exact solution of the ODE system,
593     \begin{equation}
594     \frac{{dx}}{{dt}} = f(x) \label{introEquation:ODE}
595     \end{equation}
596     The exact flow(solution) $\varphi_\tau$ is defined by
597     \[
598     x(t+\tau) =\varphi_\tau(x(t))
599     \]
600     where $\tau$ is a fixed time step and $\varphi$ is a map from phase
601 tim 2702 space to itself. The flow has the continuous group property,
602 tim 2698 \begin{equation}
603 tim 2702 \varphi _{\tau _1 } \circ \varphi _{\tau _2 } = \varphi _{\tau _1
604     + \tau _2 } .
605     \end{equation}
606     In particular,
607     \begin{equation}
608     \varphi _\tau \circ \varphi _{ - \tau } = I
609     \end{equation}
610     Therefore, the exact flow is self-adjoint,
611     \begin{equation}
612     \varphi _\tau = \varphi _{ - \tau }^{ - 1}.
613     \end{equation}
614     The exact flow can also be written in terms of the of an operator,
615     \begin{equation}
616     \varphi _\tau (x) = e^{\tau \sum\limits_i {f_i (x)\frac{\partial
617     }{{\partial x_i }}} } (x) \equiv \exp (\tau f)(x).
618     \label{introEquation:exponentialOperator}
619     \end{equation}
620    
621     In most cases, it is not easy to find the exact flow $\varphi_\tau$.
622     Instead, we use a approximate map, $\psi_\tau$, which is usually
623     called integrator. The order of an integrator $\psi_\tau$ is $p$, if
624     the Taylor series of $\psi_\tau$ agree to order $p$,
625     \begin{equation}
626 tim 2698 \psi_tau(x) = x + \tau f(x) + O(\tau^{p+1})
627     \end{equation}
628    
629 tim 2702 \subsection{\label{introSection:geometricProperties}Geometric Properties}
630    
631 tim 2698 The hidden geometric properties of ODE and its flow play important
632 tim 2702 roles in numerical studies. Many of them can be found in systems
633     which occur naturally in applications.
634    
635     Let $\varphi$ be the flow of Hamiltonian vector field, $\varphi$ is
636     a \emph{symplectic} flow if it satisfies,
637 tim 2698 \begin{equation}
638 tim 2703 {\varphi '}^T J \varphi ' = J.
639 tim 2698 \end{equation}
640     According to Liouville's theorem, the symplectic volume is invariant
641     under a Hamiltonian flow, which is the basis for classical
642 tim 2699 statistical mechanics. Furthermore, the flow of a Hamiltonian vector
643     field on a symplectic manifold can be shown to be a
644     symplectomorphism. As to the Poisson system,
645 tim 2698 \begin{equation}
646 tim 2703 {\varphi '}^T J \varphi ' = J \circ \varphi
647 tim 2698 \end{equation}
648 tim 2702 is the property must be preserved by the integrator.
649    
650     It is possible to construct a \emph{volume-preserving} flow for a
651     source free($ \nabla \cdot f = 0 $) ODE, if the flow satisfies $
652     \det d\varphi = 1$. One can show easily that a symplectic flow will
653     be volume-preserving.
654    
655     Changing the variables $y = h(x)$ in a ODE\ref{introEquation:ODE}
656     will result in a new system,
657 tim 2698 \[
658     \dot y = \tilde f(y) = ((dh \cdot f)h^{ - 1} )(y).
659     \]
660     The vector filed $f$ has reversing symmetry $h$ if $f = - \tilde f$.
661     In other words, the flow of this vector field is reversible if and
662 tim 2702 only if $ h \circ \varphi ^{ - 1} = \varphi \circ h $.
663 tim 2698
664 tim 2702 When designing any numerical methods, one should always try to
665     preserve the structural properties of the original ODE and its flow.
666    
667 tim 2699 \subsection{\label{introSection:constructionSymplectic}Construction of Symplectic Methods}
668     A lot of well established and very effective numerical methods have
669     been successful precisely because of their symplecticities even
670     though this fact was not recognized when they were first
671     constructed. The most famous example is leapfrog methods in
672     molecular dynamics. In general, symplectic integrators can be
673     constructed using one of four different methods.
674     \begin{enumerate}
675     \item Generating functions
676     \item Variational methods
677     \item Runge-Kutta methods
678     \item Splitting methods
679     \end{enumerate}
680 tim 2698
681 tim 2699 Generating function tends to lead to methods which are cumbersome
682 tim 2702 and difficult to use. In dissipative systems, variational methods
683     can capture the decay of energy accurately. Since their
684     geometrically unstable nature against non-Hamiltonian perturbations,
685     ordinary implicit Runge-Kutta methods are not suitable for
686     Hamiltonian system. Recently, various high-order explicit
687     Runge--Kutta methods have been developed to overcome this
688 tim 2703 instability. However, due to computational penalty involved in
689     implementing the Runge-Kutta methods, they do not attract too much
690     attention from Molecular Dynamics community. Instead, splitting have
691     been widely accepted since they exploit natural decompositions of
692     the system\cite{Tuckerman92}.
693 tim 2702
694     \subsubsection{\label{introSection:splittingMethod}Splitting Method}
695    
696     The main idea behind splitting methods is to decompose the discrete
697     $\varphi_h$ as a composition of simpler flows,
698 tim 2699 \begin{equation}
699     \varphi _h = \varphi _{h_1 } \circ \varphi _{h_2 } \ldots \circ
700     \varphi _{h_n }
701     \label{introEquation:FlowDecomposition}
702     \end{equation}
703     where each of the sub-flow is chosen such that each represent a
704 tim 2702 simpler integration of the system.
705    
706     Suppose that a Hamiltonian system takes the form,
707     \[
708     H = H_1 + H_2.
709     \]
710     Here, $H_1$ and $H_2$ may represent different physical processes of
711     the system. For instance, they may relate to kinetic and potential
712     energy respectively, which is a natural decomposition of the
713     problem. If $H_1$ and $H_2$ can be integrated using exact flows
714     $\varphi_1(t)$ and $\varphi_2(t)$, respectively, a simple first
715     order is then given by the Lie-Trotter formula
716 tim 2699 \begin{equation}
717 tim 2702 \varphi _h = \varphi _{1,h} \circ \varphi _{2,h},
718     \label{introEquation:firstOrderSplitting}
719     \end{equation}
720     where $\varphi _h$ is the result of applying the corresponding
721     continuous $\varphi _i$ over a time $h$. By definition, as
722     $\varphi_i(t)$ is the exact solution of a Hamiltonian system, it
723     must follow that each operator $\varphi_i(t)$ is a symplectic map.
724     It is easy to show that any composition of symplectic flows yields a
725     symplectic map,
726     \begin{equation}
727 tim 2699 (\varphi '\phi ')^T J\varphi '\phi ' = \phi '^T \varphi '^T J\varphi
728 tim 2702 '\phi ' = \phi '^T J\phi ' = J,
729 tim 2699 \label{introEquation:SymplecticFlowComposition}
730     \end{equation}
731 tim 2702 where $\phi$ and $\psi$ both are symplectic maps. Thus operator
732     splitting in this context automatically generates a symplectic map.
733 tim 2699
734 tim 2702 The Lie-Trotter splitting(\ref{introEquation:firstOrderSplitting})
735     introduces local errors proportional to $h^2$, while Strang
736     splitting gives a second-order decomposition,
737     \begin{equation}
738     \varphi _h = \varphi _{1,h/2} \circ \varphi _{2,h} \circ \varphi
739     _{1,h/2} ,
740     \label{introEqaution:secondOrderSplitting}
741     \end{equation}
742     which has a local error proportional to $h^3$. Sprang splitting's
743     popularity in molecular simulation community attribute to its
744     symmetric property,
745     \begin{equation}
746     \varphi _h^{ - 1} = \varphi _{ - h}.
747 tim 2703 \label{introEquation:timeReversible}
748 tim 2702 \end{equation}
749    
750     \subsubsection{\label{introSection:exampleSplittingMethod}Example of Splitting Method}
751     The classical equation for a system consisting of interacting
752     particles can be written in Hamiltonian form,
753     \[
754     H = T + V
755     \]
756     where $T$ is the kinetic energy and $V$ is the potential energy.
757     Setting $H_1 = T, H_2 = V$ and applying Strang splitting, one
758     obtains the following:
759     \begin{align}
760     q(\Delta t) &= q(0) + \dot{q}(0)\Delta t +
761     \frac{F[q(0)]}{m}\frac{\Delta t^2}{2}, %
762     \label{introEquation:Lp10a} \\%
763     %
764     \dot{q}(\Delta t) &= \dot{q}(0) + \frac{\Delta t}{2m}
765     \biggl [F[q(0)] + F[q(\Delta t)] \biggr]. %
766     \label{introEquation:Lp10b}
767     \end{align}
768     where $F(t)$ is the force at time $t$. This integration scheme is
769     known as \emph{velocity verlet} which is
770     symplectic(\ref{introEquation:SymplecticFlowComposition}),
771     time-reversible(\ref{introEquation:timeReversible}) and
772     volume-preserving (\ref{introEquation:volumePreserving}). These
773     geometric properties attribute to its long-time stability and its
774     popularity in the community. However, the most commonly used
775     velocity verlet integration scheme is written as below,
776     \begin{align}
777     \dot{q}\biggl (\frac{\Delta t}{2}\biggr ) &=
778     \dot{q}(0) + \frac{\Delta t}{2m}\, F[q(0)], \label{introEquation:Lp9a}\\%
779     %
780     q(\Delta t) &= q(0) + \Delta t\, \dot{q}\biggl (\frac{\Delta t}{2}\biggr ),%
781     \label{introEquation:Lp9b}\\%
782     %
783     \dot{q}(\Delta t) &= \dot{q}\biggl (\frac{\Delta t}{2}\biggr ) +
784     \frac{\Delta t}{2m}\, F[q(0)]. \label{introEquation:Lp9c}
785     \end{align}
786     From the preceding splitting, one can see that the integration of
787     the equations of motion would follow:
788     \begin{enumerate}
789     \item calculate the velocities at the half step, $\frac{\Delta t}{2}$, from the forces calculated at the initial position.
790    
791     \item Use the half step velocities to move positions one whole step, $\Delta t$.
792    
793     \item Evaluate the forces at the new positions, $\mathbf{r}(\Delta t)$, and use the new forces to complete the velocity move.
794    
795     \item Repeat from step 1 with the new position, velocities, and forces assuming the roles of the initial values.
796     \end{enumerate}
797    
798     Simply switching the order of splitting and composing, a new
799     integrator, the \emph{position verlet} integrator, can be generated,
800     \begin{align}
801     \dot q(\Delta t) &= \dot q(0) + \Delta tF(q(0))\left[ {q(0) +
802     \frac{{\Delta t}}{{2m}}\dot q(0)} \right], %
803     \label{introEquation:positionVerlet1} \\%
804     %
805 tim 2703 q(\Delta t) &= q(0) + \frac{{\Delta t}}{2}\left[ {\dot q(0) + \dot
806 tim 2702 q(\Delta t)} \right]. %
807     \label{introEquation:positionVerlet1}
808     \end{align}
809    
810     \subsubsection{\label{introSection:errorAnalysis}Error Analysis and Higher Order Methods}
811    
812     Baker-Campbell-Hausdorff formula can be used to determine the local
813     error of splitting method in terms of commutator of the
814     operators(\ref{introEquation:exponentialOperator}) associated with
815     the sub-flow. For operators $hX$ and $hY$ which are associate to
816     $\varphi_1(t)$ and $\varphi_2(t$ respectively , we have
817     \begin{equation}
818     \exp (hX + hY) = \exp (hZ)
819     \end{equation}
820     where
821     \begin{equation}
822     hZ = hX + hY + \frac{{h^2 }}{2}[X,Y] + \frac{{h^3 }}{2}\left(
823     {[X,[X,Y]] + [Y,[Y,X]]} \right) + \ldots .
824     \end{equation}
825     Here, $[X,Y]$ is the commutators of operator $X$ and $Y$ given by
826     \[
827     [X,Y] = XY - YX .
828     \]
829     Applying Baker-Campbell-Hausdorff formula to Sprang splitting, we
830     can obtain
831 tim 2703 \begin{eqnarray*}
832 tim 2702 \exp (h X/2)\exp (h Y)\exp (h X/2) & = & \exp (h X + h Y + h^2
833 tim 2703 [X,Y]/4 + h^2 [Y,X]/4 \\ & & \mbox{} + h^2 [X,X]/8 + h^2 [Y,Y]/8 \\
834     & & \mbox{} + h^3 [Y,[Y,X]]/12 - h^3 [X,[X,Y]]/24 & & \mbox{} +
835     \ldots )
836     \end{eqnarray*}
837 tim 2702 Since \[ [X,Y] + [Y,X] = 0\] and \[ [X,X] = 0\], the dominant local
838     error of Spring splitting is proportional to $h^3$. The same
839     procedure can be applied to general splitting, of the form
840     \begin{equation}
841     \varphi _{b_m h}^2 \circ \varphi _{a_m h}^1 \circ \varphi _{b_{m -
842     1} h}^2 \circ \ldots \circ \varphi _{a_1 h}^1 .
843     \end{equation}
844     Careful choice of coefficient $a_1 ,\ldot , b_m$ will lead to higher
845     order method. Yoshida proposed an elegant way to compose higher
846     order methods based on symmetric splitting. Given a symmetric second
847     order base method $ \varphi _h^{(2)} $, a fourth-order symmetric
848     method can be constructed by composing,
849     \[
850     \varphi _h^{(4)} = \varphi _{\alpha h}^{(2)} \circ \varphi _{\beta
851     h}^{(2)} \circ \varphi _{\alpha h}^{(2)}
852     \]
853     where $ \alpha = - \frac{{2^{1/3} }}{{2 - 2^{1/3} }}$ and $ \beta
854     = \frac{{2^{1/3} }}{{2 - 2^{1/3} }}$. Moreover, a symmetric
855     integrator $ \varphi _h^{(2n + 2)}$ can be composed by
856     \begin{equation}
857     \varphi _h^{(2n + 2)} = \varphi _{\alpha h}^{(2n)} \circ \varphi
858     _{\beta h}^{(2n)} \circ \varphi _{\alpha h}^{(2n)}
859     \end{equation}
860     , if the weights are chosen as
861     \[
862     \alpha = - \frac{{2^{1/(2n + 1)} }}{{2 - 2^{1/(2n + 1)} }},\beta =
863     \frac{{2^{1/(2n + 1)} }}{{2 - 2^{1/(2n + 1)} }} .
864     \]
865    
866 tim 2694 \section{\label{introSection:molecularDynamics}Molecular Dynamics}
867    
868     As a special discipline of molecular modeling, Molecular dynamics
869     has proven to be a powerful tool for studying the functions of
870     biological systems, providing structural, thermodynamic and
871     dynamical information.
872    
873     \subsection{\label{introSec:mdInit}Initialization}
874    
875     \subsection{\label{introSection:mdIntegration} Integration of the Equations of Motion}
876    
877 tim 2693 \section{\label{introSection:rigidBody}Dynamics of Rigid Bodies}
878 tim 2692
879 tim 2694 A rigid body is a body in which the distance between any two given
880     points of a rigid body remains constant regardless of external
881     forces exerted on it. A rigid body therefore conserves its shape
882     during its motion.
883    
884     Applications of dynamics of rigid bodies.
885    
886 tim 2695 \subsection{\label{introSection:lieAlgebra}Lie Algebra}
887 tim 2694
888 tim 2695 \subsection{\label{introSection:DLMMotionEquation}The Euler Equations of Rigid Body Motion}
889    
890     \subsection{\label{introSection:otherRBMotionEquation}Other Formulations for Rigid Body Motion}
891    
892 tim 2693 \section{\label{introSection:correlationFunctions}Correlation Functions}
893 tim 2692
894 tim 2685 \section{\label{introSection:langevinDynamics}Langevin Dynamics}
895    
896 tim 2696 \subsection{\label{introSection:LDIntroduction}Introduction and application of Langevin Dynamics}
897    
898 tim 2692 \subsection{\label{introSection:generalizedLangevinDynamics}Generalized Langevin Dynamics}
899 tim 2685
900 tim 2696 \begin{equation}
901     H = \frac{{p^2 }}{{2m}} + U(x) + H_B + \Delta U(x,x_1 , \ldots x_N)
902     \label{introEquation:bathGLE}
903     \end{equation}
904     where $H_B$ is harmonic bath Hamiltonian,
905     \[
906     H_B =\sum\limits_{\alpha = 1}^N {\left\{ {\frac{{p_\alpha ^2
907     }}{{2m_\alpha }} + \frac{1}{2}m_\alpha w_\alpha ^2 } \right\}}
908     \]
909     and $\Delta U$ is bilinear system-bath coupling,
910     \[
911     \Delta U = - \sum\limits_{\alpha = 1}^N {g_\alpha x_\alpha x}
912     \]
913     Completing the square,
914     \[
915     H_B + \Delta U = \sum\limits_{\alpha = 1}^N {\left\{
916     {\frac{{p_\alpha ^2 }}{{2m_\alpha }} + \frac{1}{2}m_\alpha
917     w_\alpha ^2 \left( {x_\alpha - \frac{{g_\alpha }}{{m_\alpha
918     w_\alpha ^2 }}x} \right)^2 } \right\}} - \sum\limits_{\alpha =
919     1}^N {\frac{{g_\alpha ^2 }}{{2m_\alpha w_\alpha ^2 }}} x^2
920     \]
921     and putting it back into Eq.~\ref{introEquation:bathGLE},
922     \[
923     H = \frac{{p^2 }}{{2m}} + W(x) + \sum\limits_{\alpha = 1}^N
924     {\left\{ {\frac{{p_\alpha ^2 }}{{2m_\alpha }} + \frac{1}{2}m_\alpha
925     w_\alpha ^2 \left( {x_\alpha - \frac{{g_\alpha }}{{m_\alpha
926     w_\alpha ^2 }}x} \right)^2 } \right\}}
927     \]
928     where
929     \[
930     W(x) = U(x) - \sum\limits_{\alpha = 1}^N {\frac{{g_\alpha ^2
931     }}{{2m_\alpha w_\alpha ^2 }}} x^2
932     \]
933     Since the first two terms of the new Hamiltonian depend only on the
934     system coordinates, we can get the equations of motion for
935     Generalized Langevin Dynamics by Hamilton's equations
936     \ref{introEquation:motionHamiltonianCoordinate,
937     introEquation:motionHamiltonianMomentum},
938     \begin{align}
939     \dot p &= - \frac{{\partial H}}{{\partial x}}
940     &= m\ddot x
941     &= - \frac{{\partial W(x)}}{{\partial x}} - \sum\limits_{\alpha = 1}^N {g_\alpha \left( {x_\alpha - \frac{{g_\alpha }}{{m_\alpha w_\alpha ^2 }}x} \right)}
942 tim 2702 \label{introEquation:Lp5}
943 tim 2696 \end{align}
944     , and
945     \begin{align}
946     \dot p_\alpha &= - \frac{{\partial H}}{{\partial x_\alpha }}
947     &= m\ddot x_\alpha
948     &= \- m_\alpha w_\alpha ^2 \left( {x_\alpha - \frac{{g_\alpha}}{{m_\alpha w_\alpha ^2 }}x} \right)
949     \end{align}
950    
951     \subsection{\label{introSection:laplaceTransform}The Laplace Transform}
952    
953     \[
954     L(x) = \int_0^\infty {x(t)e^{ - pt} dt}
955     \]
956    
957     \[
958     L(x + y) = L(x) + L(y)
959     \]
960    
961     \[
962     L(ax) = aL(x)
963     \]
964    
965     \[
966     L(\dot x) = pL(x) - px(0)
967     \]
968    
969     \[
970     L(\ddot x) = p^2 L(x) - px(0) - \dot x(0)
971     \]
972    
973     \[
974     L\left( {\int_0^t {g(t - \tau )h(\tau )d\tau } } \right) = G(p)H(p)
975     \]
976    
977     Some relatively important transformation,
978     \[
979     L(\cos at) = \frac{p}{{p^2 + a^2 }}
980     \]
981    
982     \[
983     L(\sin at) = \frac{a}{{p^2 + a^2 }}
984     \]
985    
986     \[
987     L(1) = \frac{1}{p}
988     \]
989    
990     First, the bath coordinates,
991     \[
992     p^2 L(x_\alpha ) - px_\alpha (0) - \dot x_\alpha (0) = - \omega
993     _\alpha ^2 L(x_\alpha ) + \frac{{g_\alpha }}{{\omega _\alpha
994     }}L(x)
995     \]
996     \[
997     L(x_\alpha ) = \frac{{\frac{{g_\alpha }}{{\omega _\alpha }}L(x) +
998     px_\alpha (0) + \dot x_\alpha (0)}}{{p^2 + \omega _\alpha ^2 }}
999     \]
1000     Then, the system coordinates,
1001     \begin{align}
1002     mL(\ddot x) &= - \frac{1}{p}\frac{{\partial W(x)}}{{\partial x}} -
1003     \sum\limits_{\alpha = 1}^N {\left\{ {\frac{{\frac{{g_\alpha
1004     }}{{\omega _\alpha }}L(x) + px_\alpha (0) + \dot x_\alpha
1005     (0)}}{{p^2 + \omega _\alpha ^2 }} - \frac{{g_\alpha ^2 }}{{m_\alpha
1006     }}\omega _\alpha ^2 L(x)} \right\}}
1007     %
1008     &= - \frac{1}{p}\frac{{\partial W(x)}}{{\partial x}} -
1009     \sum\limits_{\alpha = 1}^N {\left\{ { - \frac{{g_\alpha ^2 }}{{m_\alpha \omega _\alpha ^2 }}\frac{p}{{p^2 + \omega _\alpha ^2 }}pL(x)
1010     - \frac{p}{{p^2 + \omega _\alpha ^2 }}g_\alpha x_\alpha (0)
1011     - \frac{1}{{p^2 + \omega _\alpha ^2 }}g_\alpha \dot x_\alpha (0)} \right\}}
1012     \end{align}
1013     Then, the inverse transform,
1014    
1015     \begin{align}
1016     m\ddot x &= - \frac{{\partial W(x)}}{{\partial x}} -
1017     \sum\limits_{\alpha = 1}^N {\left\{ {\left( { - \frac{{g_\alpha ^2
1018     }}{{m_\alpha \omega _\alpha ^2 }}} \right)\int_0^t {\cos (\omega
1019     _\alpha t)\dot x(t - \tau )d\tau - \left[ {g_\alpha x_\alpha (0)
1020     - \frac{{g_\alpha }}{{m_\alpha \omega _\alpha }}} \right]\cos
1021     (\omega _\alpha t) - \frac{{g_\alpha \dot x_\alpha (0)}}{{\omega
1022     _\alpha }}\sin (\omega _\alpha t)} } \right\}}
1023     %
1024     &= - \frac{{\partial W(x)}}{{\partial x}} - \int_0^t
1025     {\sum\limits_{\alpha = 1}^N {\left( { - \frac{{g_\alpha ^2
1026     }}{{m_\alpha \omega _\alpha ^2 }}} \right)\cos (\omega _\alpha
1027     t)\dot x(t - \tau )d} \tau } + \sum\limits_{\alpha = 1}^N {\left\{
1028     {\left[ {g_\alpha x_\alpha (0) - \frac{{g_\alpha }}{{m_\alpha
1029     \omega _\alpha }}} \right]\cos (\omega _\alpha t) +
1030     \frac{{g_\alpha \dot x_\alpha (0)}}{{\omega _\alpha }}\sin
1031     (\omega _\alpha t)} \right\}}
1032     \end{align}
1033    
1034     \begin{equation}
1035     m\ddot x = - \frac{{\partial W}}{{\partial x}} - \int_0^t {\xi
1036     (t)\dot x(t - \tau )d\tau } + R(t)
1037     \label{introEuqation:GeneralizedLangevinDynamics}
1038     \end{equation}
1039     %where $ {\xi (t)}$ is friction kernel, $R(t)$ is random force and
1040     %$W$ is the potential of mean force. $W(x) = - kT\ln p(x)$
1041     \[
1042     \xi (t) = \sum\limits_{\alpha = 1}^N {\left( { - \frac{{g_\alpha ^2
1043     }}{{m_\alpha \omega _\alpha ^2 }}} \right)\cos (\omega _\alpha t)}
1044     \]
1045     For an infinite harmonic bath, we can use the spectral density and
1046     an integral over frequencies.
1047    
1048     \[
1049     R(t) = \sum\limits_{\alpha = 1}^N {\left( {g_\alpha x_\alpha (0)
1050     - \frac{{g_\alpha ^2 }}{{m_\alpha \omega _\alpha ^2 }}x(0)}
1051     \right)\cos (\omega _\alpha t)} + \frac{{\dot x_\alpha
1052     (0)}}{{\omega _\alpha }}\sin (\omega _\alpha t)
1053     \]
1054     The random forces depend only on initial conditions.
1055    
1056     \subsubsection{\label{introSection:secondFluctuationDissipation}The Second Fluctuation Dissipation Theorem}
1057     So we can define a new set of coordinates,
1058     \[
1059     q_\alpha (t) = x_\alpha (t) - \frac{1}{{m_\alpha \omega _\alpha
1060     ^2 }}x(0)
1061     \]
1062     This makes
1063     \[
1064     R(t) = \sum\limits_{\alpha = 1}^N {g_\alpha q_\alpha (t)}
1065     \]
1066     And since the $q$ coordinates are harmonic oscillators,
1067     \[
1068     \begin{array}{l}
1069     \left\langle {q_\alpha (t)q_\alpha (0)} \right\rangle = \left\langle {q_\alpha ^2 (0)} \right\rangle \cos (\omega _\alpha t) \\
1070     \left\langle {q_\alpha (t)q_\beta (0)} \right\rangle = \delta _{\alpha \beta } \left\langle {q_\alpha (t)q_\alpha (0)} \right\rangle \\
1071     \end{array}
1072     \]
1073    
1074     \begin{align}
1075     \left\langle {R(t)R(0)} \right\rangle &= \sum\limits_\alpha
1076     {\sum\limits_\beta {g_\alpha g_\beta \left\langle {q_\alpha
1077     (t)q_\beta (0)} \right\rangle } }
1078     %
1079     &= \sum\limits_\alpha {g_\alpha ^2 \left\langle {q_\alpha ^2 (0)}
1080     \right\rangle \cos (\omega _\alpha t)}
1081     %
1082     &= kT\xi (t)
1083     \end{align}
1084    
1085     \begin{equation}
1086     \xi (t) = \left\langle {R(t)R(0)} \right\rangle
1087     \label{introEquation:secondFluctuationDissipation}
1088     \end{equation}
1089    
1090     \section{\label{introSection:hydroynamics}Hydrodynamics}
1091    
1092     \subsection{\label{introSection:frictionTensor} Friction Tensor}
1093     \subsection{\label{introSection:analyticalApproach}Analytical
1094     Approach}
1095    
1096     \subsection{\label{introSection:approximationApproach}Approximation
1097     Approach}
1098    
1099     \subsection{\label{introSection:centersRigidBody}Centers of Rigid
1100     Body}