Skip to content

Latest commit

 

History

History
296 lines (213 loc) · 12 KB

README.md

File metadata and controls

296 lines (213 loc) · 12 KB
Raw

SSJ with tools for network flow problems (under construction)

IMPORTANT: This is NOT a new version of SSJ. This repository contains an old version of SSJ from 2018, together with several classes that are under construction (incomplete), to handle graphs and networks and implement various reliability and flow algorithms. It was forked from shmodal (Sofiane Hadji).

The rest of this README file (below) was copied directly from the SSJ repository and is not specific to the present one.

=========================================================================

Stochastic Simulation in Java

SSJ is a Java library for stochastic simulation, developed under the supervision of Pierre L'Ecuyer in the Simulation and Optimization Laboratory, Department of Computer Science and Operations Research at Université de Montréal. It provides facilities for:

  • random number and random variate generation
  • stochastic process simulation
  • discrete-event simulation
  • computations with several types of probability distributions
  • randomized quasi-Monte Carlo methods
  • collecting and reporting statistics from simulations
  • goodness-of-fit tests
  • and much more.

Starting from version 3.1.0, SSJ is released under the Apache 2.0 License, and the package names have changed from umontreal.iro.lecuyer.* to umontreal.ssj.*.

Documentation and tutorial

The SSJ User's Guide includes:

Installation

You can install SSJ either by adding it as a dependency for your Maven- or Gradle-based project, by downloading a binary release or by compiling it from scratch.

SSJ is compatible with Java SE8 (JDK 8) and later versions of Java. The latest Java JDK is available at Oracle with installation instructions. It should be installed before installing SSJ.

Using Maven

(Simplest approach)

SSJ packages are hosted on Bintray and on Maven Central. If your Java project uses Maven or Gradle, all you need to do is add ca.umontreal.iro.simul:ssj:+ to the Maven dependencies of your project, then you can start working on your SSJ-based project right-away.

IDE integration

Several integrated development environments (IDEs) such as Eclipse, NetBeans, IntelliJ IDEA, for example, support Maven.
If you work in one of these IDEs, simply create your project as a Maven project instead of a Java project, then add SSJ to the Maven dependencies of the project (refer to your IDE documentation), with the following parameters:

and you are ready to go!

Binary releases

For those who want to download the binaries and install them manually, we provide below some general instructions for configuring a project to use SSJ. Less experienced users can find more detailed instructions on the SSJ page.

Download a binary archive

Pre-compiled binaries are available as archives on the releases page. They include:

  • the main SSJ JAR file (under ssj/lib);
  • JAR files for the Java libraries used by SSJ (the dependencies) (under ssj/lib);
  • dependencies and the JNI shared libraries (under the ssj/lib directory);
  • the user's guide (under ssj/doc/html); and
  • example source files (under ssj/doc/examples).

You can download the latest archive and extract the files in a location of your choice.

Set the Java class path

You add to the Java class path every JAR file found under the ssj/lib directory of the binary archive.

On the command line

If you use Java from the command line, add the full path of every JAR file under ssj/lib to the CLASSPATH environment variable, separated with : under Linux or MacOS, or by ; under Windows. Means of doing this depends on the system you are using. For example, under Linux with a Bash-compatible shell, one could use something like:

for f in /full/path/to/ssj/lib/*.jar; do
    CLASSPATH=$f:$CLASSPATH
done
export CLASSPATH
In Eclipse

In Eclipse, under Window ‣ Preferences ‣ Java ‣ Build Path ‣ User Libraries, click New…. Set the name to SSJ and click OK. Click Add External JARs… navigate to the ssj/lib folder of the extracted binary archive, select all JAR files, and click OK. You can now add the SSJ library you have created from any project, by right-clicking on your project name in the Package Explorer, by selecting Build Path ‣ Add Libraries… ‣ User Library under your project tree and by choosing SSJ.

In NetBeans

In NetBeans, under Tools ‣ Libraries, press New Library…. Set the name to SSJ and click OK. Click Add JAR/Folder…, navigate to the ssj/lib folder of the extracted binary archive, select all JAR files, and click Add JAR/Folder. You can now add the SSJ library you have created from any project, by right-clicking on Libraries under your project tree in the Projects tab and by choosing SSJ.

Compiling the source code

You do not need to compile the source code to use SSJ if you have already installed it using Maven or a binary release. But in case you want to change the source for some reason, here is how you can recompile.

The SSJ library uses Gradle as its build system. You do not need to download it, since the Gradle wrapper executable program is provided with the source code as the gradlew file for Linux and MacOS platforms, and as gradlew.bat for Windows platforms. The build.gradle and gradle.properties files at the root of the source tree contain the configuration for Gradle. In the instructions below, Windows users should replace instances of ./gradlew with gradlew.bat.

SSJ and the current Gradle version work with Java SE (or JDK) version 8 or later.

Using Gradle

The general syntax for Gradle is ./gradlew <task> where <task> is the name of a Gradle task (run ./gradlew tasks to obtain a list of available tasks).

On the command line, from the root of the source tree, type:

  • ./gradlew check to build and test the library;
  • ./gradlew examples (optionally) to run additional examples;
  • ./gradlew distZip or ./gradlew distTar to create an binary archive of the SSJ library, including the SSJ JAR file and its dependencies;
  • ./gradlew --gui to launch the Gradle graphical user interface and choose from more options.

All files generated during the build process are placed under the build subdirectory. The generated binary archives can be found under build/distributions.

Building the documentation

(Optional)

The SSJ library uses Doxygen as its documentation system. If Doxygen is available on your system, you can tell Gradle to build the documentation by adding the following line in gradle.properties:

buildDocs

Then, run Gradle as explained above. You may want (or need) to change some Doxygen environments variables in the file Doxyfile.

JNI classes

(Optional)

The classes UnuranContinuous, UnuranDiscreteInt, UnuranEmpirical and GlobalCPUTimeChrono make use of native libraries through the Java Native Interface (JNI). These libraries must be compiled with a C compiler (known to work with GCC).

Note that if you want to build and use the UNU.RAN interface provided with SSJ, you must first install UNU.RAN.

To tell Gradle to build the JNI libraries, add the following lines in gradle.properties:

ssjutil.jni.build
randvar.jni.build
unuran.prefix = "/path/to/unuran"

Make sure to replace /path/to/unuran in the above with the installation prefix of UNU.RAN, i.e., the path under which include/unuran.h can be found. Then, run Gradle as explained above.

Cross-compiling

(For experts only!)

Under Linux with GCC and MinGW, you can cross-compile the JNI libraries for both Linux and Windows. This is how we generate the binary archives.

To enable cross-compilation with Gradle, add the following lines in gradle.properties:

crossCompile
unuran.prefix.linux64 = "/path/to/unuran"
unuran.prefix.win32   = "/path/to/unuran-mingw32"

Make sure to replace /path/to/unuran and /path/to/unuran-mingw32 with the installation prefixes of UNU.RAN compiled for 64-bit Linux and for 32-bit Windows, respectively.

Dependencies

SSJ depends on the following libraries. You do not need to download them manually if you're using SSJ with Maven or from a binary release.

If you intend to compile the source code of SSJ, Gradle will take care of downloading the Java dependencies for you. Optionally, if you want to use the UNU.RAN interface in SSJ, you need to install the UNU.RAN before compiling the associated JNI shared library in SSJ.

Colt

The Colt library is used by a few SSJ classes. The library, its source code and documentation, can be downloaded for free from its home page. The colt.jar archive is already included in the SSJ distribution and it must be in the CLASSPATH environment variable.

Nonlinear Optimization Java Package by Steve Verrill

The optimization package of Steve Verrill includes Java translations of the MINPACK nonlinear least squares routines as well as UNCMIN routines for unconstrained optimization. They were translated from FORTRAN to Java by Steve Verrill and are in the public domain. They are included in the SSJ distribution as the optimization.jar archive. It is used only in the probdist package to compute maximum likelihood estimators.

JFreeChart

The JFreeChart library is used by the SSJ package charts to draw curves, histograms and different kinds of plots. JFreeChart is copyrighted under the GNU LGPL License. It is included in the SSJ distribution as the jfreechart-X.Y.Z.jar and the jcommon-X.Y.Z.jar, where X.Y.Z represents a version number.

UNU.RAN (optional)

The UNURAN library is used by the classes UnuranContinuous, UnuranDiscrete and UnuranEmpirical in the package called randvar. Downloading, compiling and installing UNURAN is optional. It is required only if SSJ must be rebuilt. However, the UNURAN documentation is required to use the SSJ UNURAN interface efficiently.