Research on Multi-Agent Task Optimization and Scheduling Based on Improved Ant Colony Algorithm

Abstract

Abstract As a key problem of cloud computing, the performance of task scheduling strategy may seriously affect the efficiency and service quality of the system. With the purpose of achieving balanced task scheduling with optimal completion time, ant colony optimization (ACO) algorithm is adopted as task scheduling strategy in this paper. Considering the problem of premature convergence to local optimal solution of ACO, an improved Max-Min Ant System (MMAS) is introduced to find the global optimal solution. Based on the study of pheromone’s updating mechanism of MMAS, MMAS is applied as task scheduling strategy with reasonable mapping from the objective of shortest path to shortest completion time. The task scheduling strategy based on MMAS has been simulated from views of tasks’ number, size and load balance, and the results shows that MMAS task scheduling strategy is with a better performance on completion time and load balance than ACO based strategy.?