mh4j / metaheuristics4java Goto Github PK

The test LocalSearchSorterTest#testCreateRandomNeighbor fails if the following seeds are used:

• 1337144491848 for LocalSearchSorter
• 24205843415903631 for Sorting.createRandomSorting

This is because the neighbor created by LocalSearchSorter#createRandomNeighbor can be the current solution again, but the test tests for inequality of the current solution an the created neighbor.

Jenkins should automatically upload the latest binaries and sources of the develop-branch as file downloads to GitHub.
It would also be nice if there are artifacts including all dependencies. Looks like this is a place to start.

It would be very useful to create a wiki article that explains the default choice for the acceptance probability.
Some references and a sound foundation of the theoretical implications and possible drawbacks or traps for the user would be also very nice.

I'm not even sure yet if the current implementation is correct

We need to choose and provide an open-source license with our code.

Create wiki entry for LocalSearchSorter and SimulatedAnnealingSorter examples

A nice wiki example page or tutorial could help future users to get started with the genetic algorithm implementation

Integrate legacy Genetic Algorithms Code into the repository.

A nice wiki example page or tutorial could help future users to get started with the genetic algorithm implementation.

Adapt genetic algorithms to solve sorting.

Create a website based on the readme. Perhaps using Github Pages with Jekyll.

Write README containing the following topics

• main goals / abstract
• installation instructions
• current features
• planned features

Note If the file 'README' has the suffix '.md' you can use github markdown for formatting purposes.

Adapt simulated annealing to solve sorting & create wiki entry.

Integrate legacy Simulated Annealing Code into the repository.

We need to update this to a better function that applies to most of the possible problems and is more robust than the old implementation (log function ?!)

It would be nice if one could specify a maximum time that a simulation may run.
Absolute and relative time values should both be possible:

Examples:

• optimization can run at most 60 seconds
• optimization can run at most 1 hour
• optimization can run at most 2 days
• optimization can run at most 1.5 weeks
• optimization should not run no later than today 19:00
• optimization should not run no later than 03.07.2012

The TerminationCondition interface should be implemented.