The Study Of Multi-objective Evolutionary Algorithm For Task Scheduling In Grid

Grid computing computing has been a research focus in the field of information technology in recent years. It refers to share resources and collaborate to solve the problem in the dynamic changes heterogeneous environment. As the grid with large scale, heterogeneous, dynamic, distribution and autonomy and so on, how the most effective management and use of grid resources to the most efficient completion of a variety of computing tasks and meet the needs of users is an important of content in the grid computing, which is also called grid task scheduling problem. Grid task scheduling is one of the most important parts in grid research and is NP-hard problem. Therefore, heuristic algorithm for the purpose of obtaining the optimal solution approximately is paid great attention. So far, people propose a lot of heuristic scheduling algorithms, such as genetic algorithm and ant colony optimization and so on. Scheduling heuristic algorithm can be used to direct construction scheduling scheme. Although the task scheduling can be effective, there are also some of its own inherent flaws. These flaws are difficult to satisfy user needs to multiple conflicting performance goals of grid task scheduling at the same time. In view of the above, an adaptive neighborhood multi-objective grid task scheduling algorithm (ANMO-GTSA) is proposed in this paper for the multi-objective scheduling problems in grid computing by in-depth analysis of multi-objective optimization theory and algorithms and the principles of grid task scheduling. Moreover, this paper discusses that the radius of the neighborhood is adaptively changed by the situation of the current population, which avoids the problem that radius value of the neighborhood affected the diversity of the population in the traditional neighborhood strategy. Finialy, we adopt multi-objective evolutionary approach to solve the multiple target grid task scheduling problems in this paper.We found three main conclusions on the following by research the grid task scheduling algorithm.1. A novel multi-objective evolutionary algorithm based on adaptive neighborhood (ANMOEA) is proposed. The algorithm uses adaptive neighborhood method to maintain the distribution of population. In addition, Adaptive neighborhood radius and crowding distance density estimation are adopted to preserve the density of individuals. The experimental results indicate that the discussed method is effective in maintaining the diversity and convergence of the population, which is significantly better than NSGA-Ⅱmulti-objective algorithm.2. An adaptive neighborhood multi-objective grid task scheduling algorithm (ANMO-GTSA) is proposed. The algorithm is developed by applied ANMOEA to fast solve the synergy and balance of multiple conflicting performance objectives in grid computing task scheduling.3. In order to examine the performances of the algorithm proposed in the paper, a lot of simulation and performance analysis of the experiment are constructed by using grid scheduling simulation toolkit GridSim. The experimental results indicate that ANMO-GTSA proposed in this paper can achieve better performance than Min-min, Max-min algorithm.