16/May/2016:

• I’m moved everything under: https://github.com/stefanutti/maps-coloring-python
• Before it was on SourceForge

01/Nov/2016:

• Around this date, I started experimenting a new method based on Kempe’s chains. Read the blogs I published starting from this post

I have implemented a Python algorithm that goes really fast coloring the edges of a planar graph.

See this video and then try it youself!

Note about the Python program: To try the Python program you need to have Sage. Follow Sage instructions to install it on your computer: http://www.sagemath.org.

---

This section is dedicated to the Java program I use to analyze maps.

Snapshots of the java application:

Downloads:

To download the software and the source code of the project go to the SourceForge site:

https://github.com/stefanutti/maps-coloring-java

Youtube channel (see it in action):

http://www.youtube.com/user/mariostefanutti

Documentation:

For now (08/Feb/2011) the java software is undocumented. Best way to learn it it to start pushing buttons randomly. 😉 Don’t hesitate to contact me if you need help.

JVM settings for 1.8 (to go a great deal faster than default settings):

• java -Xms1500m -Xmx1500m -XX:+UseParallelGC -jar ct-ui-swixml-2.3-jar-with-dependencies.jar

The information that follows are not valid for the last version of the software. I implemented some code to optimize the generation of maps, such as removing isomorphic maps. For example, enabling this flag generates simplified maps. With maxMethod = F and maxNumber = 12, after filtering out maps with faces with less than 5 edges (“-5” button), it returns just one map.

Just to start coloring some good big maps follow these instructions:

• Run the application using java -jar .jar (java 1.5> and <= 1.8 is needed)
• Switch to maxMethod = F and set maxNumber = 10
• Unset the isomorphic button
• Push start and wait the elaboration to finish
• Filter out all maps that have faces with less than 4 edges (“-4” button)
• At this point you should have 14 maps left
• Browse and color them
• To generate more complex maps you can copy the maps back into the todoList (“.|.” button)
• Set maxMethod = MAPS and maxNumber = 10000
• Push play, wait the end of the elaboration and filter
• …
• filter, play and copy back into the todo list … filter, play and copy …
• …
• After some iterations you can even risk to filter maps with less than 5 edges

Another method is:

• Run the application using java -jar ct-ui-swixml-2.3-jar-with-dependencies.jar (java 1.5> and <= 1.8 is needed)
• Switch to maxMethod = F and set maxNumber = 12
• Unset the isomorphic button
• Push start and wait the elaboration to finish (less than a minute if the java GC has been set properly)
• Filter out all maps that have faces with less than 5 edges (“-5” button)
• At this point you should have 10980 maps left
• Browse and color them

Structure of the project:

• ct-core (core library for generating simplified maps)
• ct-ui-swixml (standalone UI based on swixml an G – execute using java -jar)
• …