GAMMA: A Genetic Algorithm for Multi-Module Assignment
A multiprocessor system with variable loads requires an efficient schedule for assigning modules to processors. The time it takes to execute all modules on available processors is the makespan of a schedule. The problem of finding a schedule that minimizes finish time for modules is NP-complete. Hence, heuristic methods are required to find near optimal solutions.
In this thesis, a heuristic approach is presented to schedule non identical tasks to non identical processors. The approach is to characterize processors by their relative speeds, and tasks by their relative lengths. A Genetic Algorithm is used to provide a schedule with near optimal makespan while maintaining load balance amongst processors. The details of implementation, runs and results are presented.
A multiprocessor system with variable loads requires an efficient schedule for assigning modules to processors. The time it takes to execute all modules on available processors is the…