Project Homepage - gacs.sourceforget.net
Project ScreenShots - gacs.sourceforget.net/screenshots.html

The Genetic Algorithm Class Scheduler, or GACS for short, is an open-source course scheduler based on the idea of Genetic Algorithms. The system was developed by two developers who released it under the GPL, however neither the project nor the associated website and documentation have been updated since 2003.

Good Features

• open source which allows for code customization
• requires only a web server with PHP and a SQL server installed, so it will be available to any end-user with a browser and internet connection
• provides simple interfaces to add classes and times on a weekly basis
• provides interfaces to manage and delete existing classes and times
• creates a weekly grid of the resulting schedule
• provides basic authentication for differentiating between admistrators (schedulers) and normal users (instructors)

Bad Features

• designed first and foremost as a scheduler for students; modification to the source tree would be required satisfy our specific system
• web interface, source files and directories are in Spanish
• proprietary database structure that does not adhere to our system specification (permanent course/instructor/room/preference database with per quarter copies of each)
• only two stages of operation, preliminary (creation) and final (end); no draft stages for scheduling

Missing Features

• no documentation available on the project page
• no way to optimize or tweak schedules post generation; you are stuck with the schedule that the algorithm creates
• no way to assign preferences or conflicts to instructors, courses, or rooms
• no way to re-use past schedules
• no way to view yearly schedules, just week by week
• no way to allow multiple levels of administration; just one scheduler (admin) and one end-user (normal user)
• no way to tweak the schedules after creation
• no way to assign priorities to scheduling criteria

Overall, the Genetic Algorithm Class Scheduler is not an ideal solution to our problem. The non-native English support, lack of documentation, missing features and code overhaul required to fit our spec makes this system inadequate for our purposes. However, some of the features included in this product, including the dynamic week schedule generator, data entry screens and scripts, and the algorithm behind the scheduling problem could be integrated nicely into our finished system should we choose to do this. Otherwise, the GACS is a poor choice for our system.