The Application Of Dynamic Immune Algorithm To Task Scheduling For Grid Computing

Grid Computing is a key technology which aims to achieve resource sharing and cooperative work. Task Scheduling, as an important technology for Grid Computing, its goal is to allocate tasks to grid resources by certain strategy under certain condition and make sure that total execution time reaches minimum.Firstly, we analyzed the distributed grid schedualing system and constructed a distributed scheduling model according to the characteristics of the grid by object-oriented approach. Through this step, we make clear the feature of each component of the system ,the input and output of the scheduling algorithms.Secondly, we improved the basic immune algorithm and proposed Dynamic Imunne Algorithm (DIA) to make it suitable to task scheduling for Grid Computing. For example, we redesigned the coding mechanism, fitness function, reproduction operator, crossover operator, mutation operator and the parameters of the scheduling system. By calling Dynamic Immune Algorithm, scheduling system could allocate tasks to resourses for prosessing in an optimal way considering the characteristis of the resources such as computing power and the load.Finally, we realized the scheduling model which based on DIA in GridSim.Also, we compared DIA with other common scheduling algorithms such as Genetic Algorithm, First-Come-First-Service Algorithm and Min-Min Algorithm .The results show that Dynamic Immune Algorithm is a good optimization algorithm and has self-learning ability and adaptive capacity, thus able to provide user with a better quality of service.