Task Scheduling Through a Hybrid Genetic Algorithm
Derek Seibel, Eric A. Mehiel · AIAA Modeling and Simulation Technologies Conference · 2009
An independent report has identified a plausible co ncern in the default exhaustive scheduling algorithm of the Cal Poly Horizon Simulation Framework. This concern, combined with a long-term goal of a robust and all- encompassing scheduler, has motivated the drive for an update to the Horizon scheduler. T o address this concern, a hybrid genetic algorithm is written in C++. A custom, gap-filling hybrid genetic algorithm is introduced and a phased method is demonstrated on a common benchmark system test. On average the upgraded algorithm outperformed the old scheduler 90% of the time and had a 23% faster average runtime.