Gregory R. Kramer and Frank W. Moore
The multiprocessor task scheduling problem is an NP-complete problem that is difficult to solve via traditional methods. Recent investigations have used evolutionary computation to evolve optimized schedules, but have failed to address the effects of communications delay on the schedules. As such, the schedules produced are often less than optimal. The goal of this project was to extend previous research by developing a genetic algorithm that can evolve optimized solutions to complex task scheduling problems for multiple processor systems, while taking into account the non-negligible communications delay that exists between processors.