Compile instructions

The program has been written in C++ programming language. It makes use of the CAPD library.
We merged the software package with the CAPD library snapshot available at the CAPD webpage at 2013-07-03, which has been tested to be compatible with the software package.
Preferred compiler is the GCC compiler.

The program was checked to succesfully compile with GCC compiler ver. 4.6* , 4.7.3 , 4.8.1 .

The program was tested on a Linux operating system, it is possible to compile it under Windows operating system using the MinGW environment.

Compile instructions

1. Install the CAPD library

Configuration

The configure script checks system configuration and generates Makefiles with proper parameters for compilation (from Makefile.in files). Typically it is called with one parameter

where <prefix> is localization where libraries and header files will be installed after successful compilation e.g.

Compilation and Installation

In capd directory call

• make  -  it compiles libraries, programs and examples.
• make check  -  it compiles and runs tests.
• make install  -  it copies header files, programs, libraries and files for pkg-config to <prefix> (this operation is voluntary).

Note:
The above steps can be made in some subdirectory, so that only one package or module will be compiled. The last command can require administrator rights depending on <prefix>.

If all the above steps succeeds then

• /bin will contain all programs.
• /lib will contain libraries.
• /lib/pkgconfig will contain files for pkg-config.

2. Compile the program package.

The program package does not require any extra packages of the CAPD library. It is enough to install the CAPD library in its basic version (without providing any additional parameters during configuration).

Afther the CAPD has been properly compiled and installed, the program package can be compiled and linked.

First enter the following directory within the files that were extracted from the archive capd_dynsys_dPDEs.zip

In the dPDEs directory call

3. Included programs

• FFTTests - the program, which was used to perform the overestimates tests in Section "Overestimates". Two parameters are required, first, is the number of terms in the convolution, second, defines the number of modes of a function and the number of the grid points, these numbers are prescribed in the global tables n and m
• SHTest - the program, which was used to perform the test "Swift-Hohenberg equation a connection between fixed points". Three parameters are required,
  first, is either projection or inclusion, "projection" states that the finite dimensional projection will be integrated, whereas "inclusion" states that the full infinite dimensional system will be integrated.
  Second parameter is the test index, which corresponds to the index provided in Figure 1.21.
  Third parameter is the approach type, 0 - the direct approach, 1 - the FFT approach, 2 - the FFT approach, but the first normalized derivative is calculated directly, in order to avoid blow-ups.
• KSPOTest - the program, which was used to perform the test "Kuramoto-Shivasinsky equation attracting periodic orbit".Three parameters are required,
  first, is either projection or inclusion, "projection" states that the finite dimensional projection will be integrated, whereas "inclusion" states that the full infinite dimensional system will be integrated.
  Second parameter is the test index, which corresponds to the index provided in Table 6.
  Third parameter is the approach type, 0 - the direct approach, 1 - the FFT approach, 2 - the FFT approach, but the first normalized derivative is calculated directly, in order to avoid blow-ups.
• KSUnstPOTest - the program, which was used to perform the test "Kuramoto-Shivasinsky equation unstable periodic orbit".Three parameters are required,
  first, is either projection or inclusion, "projection" states that the finite dimensional projection will be integrated, whereas "inclusion" states that the full infinite dimensional system will be integrated.
  Second parameter is the test index, which corresponds to the index provided in Table 7.
  Third parameter is the approach type, 0 - the direct approach, 1 - the FFT approach, 2 - the FFT approach, but the first normalized derivative is calculated directly, in order to avoid blow-ups.
• BurgersFPTest - the program, which was used to perform the test "Burgers equation fixed point test".Three parameters are required,
  first, has to be set to projection (is fixed), "projection" states that the finite dimensional projection will be integrated.
  Second parameter is the test index, which corresponds to the index provided in Figure 1.16.
  Third parameter is the approach type, 0 - the direct approach, 1 - the FFT approach, 2 - the FFT approach, but the first normalized derivative is calculated directly, in order to avoid blow-ups.

The programs used to compare the execution speed are not included here, because the classes implementing the FadBad++ approach has been removed from the package in order to avoid a confusion.