Visual implementation of CAA algorithm with example usage

Java implementation of the Crossbar Adaptive Array (WIP) with usage

Requirements

• Java 1.8
• Maven

Nice to have

• IntelliJ

Install

• git clone https://github.com/Tanevski3/caa-experimentation.git
• cd ./caa-experimentation
• mvn clean install

Run

• Find MainEntry.java in console or gui, then right click and run

Deploy

• cd gui
• mvn jfx:jar
• mv gui/target/jfx/app/caa-experimentation-gui-${version}-jfx.jar gui/target/jfx/app/caa-experimentation-gui-${version}.jar
• zip gui/target/jfx/app/. caa-experimentation-gui-${version}
• rm -rf dist/*
• mv caa-experimentation-gui-${version}.zip ../dist/

Usage Guide

Once "File -> Open Graph" is clicked a file selection window will appear, where the user can browse through its filesystem to select a graph for loading into the software.

In the bottom right corner just above the Open & Cancel buttons there is a list of all supported formats for loading graphs. At the time of writing this paper, GraphML format is the only format that is supported. Once a graph is loaded into the software the working space should be filled with rendering of the graph.

Changing preferences

The outlook of the graph displayed in the main area can be modified through File -> Preferences. Besides, the color of the nodes & edges within Preferences edge width & vertices labels could also be modified. Additionally, some properties specific to the experiment can be modified; that is the CAA tab.

Experimenting

Experimenting is divided in two menu-items:

1. Agent Traversal
2. Find Happy State

Experiments can be executed with different properties. Clicking on one of the menu items will show a popup for selecting properties of the algorithm used for a traversal. Properties are as follows: • Agent traversing algorithms: original & advanced • Experiment animation • Selecting initial values.

Once "Find Happy State" experiment is being ran, results will be generated in a separate tab. For agent traversal there are no results. The results relate to the graph that is being experimented on.

Experiment Results

The experiment results split in two parts.

1. Overall vertices traversals tab contains the following data a. Vertices occurrences per generation is showing how many traversals have occurred for each vertex b. "Happy" vs "Sad" generation is showing how each generation ended c. "Happy to shortest" path factor is the percentage of edges that match from the happy path to the shortest path
2. Vertices traversals per generation tab goes more intro details about a single generation. It shows the increments, traversed edges, increments weight & traversal weight. The results are being paged depending on the number of generations.

Viewing additional statistics & shortest path

Additionally, the software provides a visualization of the shortest path & vertices distribution. That can be done through the View menu.

Available as JAR

• The application is now available as a JAR file. Within the dist/ directory of this repository the executable JAR file can be found.

Check console and generated graph for results or use the application to run own simulations or traversals

Future

• Increase unit test coverage for caa-lib to 100% (current is 90%)
• Support for other formats then graphml
• Rename happy & sad to something else