This package simulates and analyzes the evolution of a more general version of the Gray-Scott reaction-diffusion model, namely one with 3 autocatalytic chemical species (i.e. a catalytic hypercycle as described in ref).
Note
This is a fork of the simulator created for my master's thesis project. The original repository is frozen to reflect the thesis and any future work will be done here.
The package was developed on Ubuntu Linux and should be relatively easy to use on other UNIX (i.e. POSIX) systems. The simulation control scripts are written in Bash but they are not necessary for compiling or running the simulator. They simply automate certain tasks.
The source code is quite portable so it shouldn't be too difficult to use on Windows with minor adjustments to things like hard-coded file paths.
-
C++
-
Python 2.7
- matplotlib >= 1.2.1
There are two simulators provided:
egs_cycle- Basic simulator in which the chemical species are defined only through the initial conditions. The simulation is not perturbed once started.egs_cycle_evol- Evolutionary simulation. The chemical species are introduced at different times to study how such perturbations affect the system at long times.
These may be compiled from src/ and run without any arguments since the
parameters are hard-coded, except for the initial conditions.
The PDE model is defined in src/include/egs_cycle_dynamics.hpp.
To compile and run the simulators, run:
egs_run.sh
and, respectively:
egs_evol_run.sh
These scripts will also run the plotting and statistics scripts (see below).
The area and initial concentration of the chemical species is specified through a
.cond file.
It is assumed that outside these specified
areas the grid holds constant [U] = 1 and [V_i] = 0 concentrations.
These files are in the cond/ subdirectory.
The .cond file is hard-coded in the simulator source code.
Simulation data is saved to the simdata/ directory and this is where the scripts
assume the data is located and where their own output is saved. On POSIX
systems, this directory may be replaced with a symbolic or hard
link should the data need saving on a larger drive.
Two scripts are included to visualize the simulation outcomes:
-
scripts/plotter.py- Plots the grid at each output generation and save these as a movie file, or, optionally, as separate frames. The script provides documentation on options. -
scripts/stats.py- Plots statistics (e.g. total concentrations over simulation time) collected from the simulation output.datfiles (see script for options).
These may be called without any arguments and the default behavior is plotting data (movie, statistics) for the entire length of the simulation.
Minimal Doxygen documentation has been generated for the simulators (start at
doc/html/index.html) though the code is meant to be self-documenting and
to be inspected/modified when in use, at least for setting parameters.
The various design decisions (like hard-coding parameters and recompiling for each run) were made to prioritize speed of execution, while keeping the architecture modular and extensible.
The program requires the GNU Scientific Library (GSL) for random number generation. GSL version 1.15 was used during development but the few functions employed should exist in much earlier version. The compiler flags to link with the library are:
-lgsl -lgslcblas -lm
Optionally, portions of the code are parallelizable with the OpenMP library, which should be included with your compiler. The program has been compiled with GCC versions 4.6.3 in the latest iteration. Earlier version should be fine, but the status of OpenMP implementation should be checked, as well as the various compiler flags. The compiler flags used were:
-O3 -Wall -Wextra -funsafe-loop-optimizations -xT -fipa-pta -ftracer -march=native -fopenmp
Only mathematically safe optimizations were enabled.
The statistics and plotting are done with Python 2.7 scripts, using matplotlib library version >= 1.2.1