Grid Task Scheduling Based On Genetic Algorithm

Grid computing is defined as"the seamless integration and cooperative computing environment"in wide regions. The grid computing pattern has developed into a cadre that connects and unifies all kinds of remote resources.Because of the characters of dynamic,distributed and different requirements to resources asked by applications running in the Grid Systems, it makes the task-scheduling very complex. A bad task-scheduling policy prolongs a task's running time and decreases the throughput of the system. Considering the complexity and essentiality of the task scheduling, it becomes an important issue in the study field of the grid computing.This dissertation suggests a grid task scheduling based on a reformative Genetic Algorithm. Considering of the differences of task scheduling between a grid and a distributed system, this dissertation designs a real number encoded mode that mapping every task to a random resource directly by improving the encoding mode of the traditional Genetic Algorithm, and gives a particular design to the encoding and decoding mode. At last, this algorithm is simulated in the Grid Simulator. The experiment results show that the reformative genetic algorithm not only has a holistic searching ability, but also makes a fast convergent speed, which provides a preferable performance. This experiment realizes the Optimal Makespan of the task scheduling...