1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
|
<html>
<head>
<title>Method Editor</title>
</head>
<body>
<h1>Method Editor</h1>
<p align="center">
<img src="MethodEditor.png">
</p>
<p>
In this panel you edit the simulation
method. (See the <a href="UserSolvers.htm">Solvers</a> section
for a description of each available solver.)
First select the time dependence category:
<ul>
<li> <tt>Time Homogeneous</tt> -
For time homogeneous problems, the reaction rates
may depend upon the species populations, but do not explicitly depend on
time. Most of Cain's solvers belong to this category. Use may choose
between exact methods (direct, next reaction, etc.) and approximate
methods (tau-leaping, hybrid, etc.).
<li> <tt>Time Inhomogeneous</tt> -
Although time inhomogeneous problems are not much more conceptually
difficult, they are more costly to simulate. Sources
of inhomogeneity may include time-varying volume or temperature (both affect
reaction rates). Note that although Gillespie's direct method is
used, the solution is not exact. We use the approximation that the
reaction propensities are assumed to be constant between reaction events.
<li> <tt>Use Events</tt> -
One must use this category if the model has events. Simulating
models with events is more costly. Additionally, the solvers in this
category are implemented in Python, instead of C++. Thus, these
solvers are significantly slower than solvers in the other
categories. Note that time inhomogeneities are allowed; the reaction
propensities, parameters, and compartment volumes may be functions
of time. However, the solutions for time inhomogeneous problems are
approximate because, again, we make the assumption that reaction
propensities are constant between reaction events.
</ul>
</p>
<p>
Next select the output category. Simulations may generate several types of
output:
<ul>
<li> <tt>Time Series, Uniform</tt> - Generate stochastic trajectories.
Record the populations and reaction counts at uniformly spaced intervals
in time.
<li> <tt>Time Series, All Reactions</tt> - Generate stochastic trajectories.
Record every reaction event. Use this output choice when you want a detailed
view of a small number of trajectories. Choose the time interval so
the number of reactions is not too large.
<li> <tt>Time Series, Deterministic</tt> - Generate a single, deterministic
trajectory by numerically integrating a set of ordinary differential
equations. Record the populations and reaction counts at uniformly spaced
intervals in time. Note that in this case the populations and reaction counts
are real numbers instead of integers.
<li> <tt>Histograms, Transient Behavior</tt> - Generate stochastic
trajectories. Record the populations in histograms at uniformly spaced
intervals in time.
Recording the species populations in histograms gives you more quantitative
information about a system than recording and plotting trajectories.
<li> <tt>Histograms, Steady State</tt> -
Record the average population values in
histograms. You may choose to start each simulation by letting the system
equilibrate for a specified time interval. This does not affect the time
during which the state is recorded. Average value histograms are useful for
determining the steady state probability distributions for the species
populations.
<li> <tt>Statistics, Transient Behavior</tt> - Record the mean and
standard deviation of the species population at each frame. There is
no solver in Cain that only records statistics. The Cain solvers for
transient behavior either record trajectories, or they record both
statistics and histograms of species populations. Thus, this acts as a
placeholder for some external solver. If one generates a solution
with another application (or records an analytical solution), then
one may import it in a text file. Each line in the file records the
statistics for a single time frame. The mean and standard deviation
for each recorded species are listed. See the
<a href="Dsmts.htm">DSMTS</a> chapter for examples.
<li> <tt>Histograms, Steady State</tt> -
Record the mean and standard deviation of the time-averaged species
populations. This is a placeholder that is used for importing
externally-generated steady state solutions. The imported text file
lists the mean and standard deviation for the recorded species on
a single line.
</ul>
</p>
<p>
In the third field select the
algorithm to generate the desired output. For each method there is a
choice of options which may affect the performance or accuracy.
</p>
<p>
In the right column, one sets the simulation parameters.
Three values control the simulation time interval: start time, equilibration
time, and recording time. The start time, as the name suggests, is the point
in time at which the simulation starts. For time homogeneous problems one
would typically choose zero. For time inhomogeneous problems one might want
to choose a nonzero value. The equilibration time is the length of time
that the simulation is advanced before recording results. One would choose a
nonzero value if one were studying the steady state behavior of a system.
For instance, a system may exhibit oscillatory behavior, but may
take a certain amount of time to establish these oscillations.
The recording time is the length of time to simulate and record the state.
</p>
<p>
Next you can select the maximum number of allowed steps when generating a
trajectory. Usually one would leave this field blank to indicate that there
is no set limit. However, this field may be useful if you don't know the
appropriate time scale for your simulation. Then you can set a limit on the
number of steps. If any of the trajectories reach this limit, the simulation
will abort with an error message.
</p>
<p>
If you have elected to record the state at frames, you choose the
number of frames to record. If not, the frames field is disabled. The
first frame is at the beginning of the recording time, and the last is
at the end. If you are only interested in the the final populations or
the final reaction counts, choose the number of frames to be one. For
this special case, the state will be recorded only at the end time. If
you are recording the output in histograms, you select the number of
bins to use in each. A histogram is an empirical probability
distribution for a species population. The computational cost of
converging these probablity distributions is related to the number of
bins. Choose a number of bins that is appropriate for the
amount of time you are willing to spend generating trajectories. Next
select the histogram multiplicity. The state is recorded in multiple
histogram arrays. This allows one to estimate the error in the
resulting distributions. Increasing the histogram multiplicity allows
one to more accurately estimate this error. However this also
increases the memory requirements for the solvers. The default value
of four is usually a reasonable compromise. If you are not interested
in estimating the error in the probability distributions,
you may set the histogram multiplicity to unity. Some
simulation methods require a parameter such as allowed error or step
size. This quantity is entered in the final box.
</p>
</body>
</html>
|
