A new task scheduling strategy based on improved ant colony algorithm in IaaS layer

In the IaaS layer of cloud computing, the cloud platform is required to implement efficient scheduling for large-scale tasks to meet the needs of both users and cloud service providers. In this paper, a multi-objective task scheduling model is established to optimize task completion time, cost and load balancing degree. Furthermore, an improved ant colony algorithm is proposed to solve this problem. The proposed algorithm firstly uses genetic operations to initialize the pheromone, and establishes an ant search model that can speed up convergence, while the pseudo-random proportional rule is used to select the next node of the path. Then, a weighted constraint function of time and cost is established to evaluate the quality of the scheduling's solution, and the Analytic Hierarchy Process (AHP) decision is used to adjust the weighting factors. Moreover, a load balancing method is introduced in the path heuristic function of ant colony algorithm. Compared with other advanced algorithms, the proposed algorithm can significantly reduce task completion time and cost, while accelerating the convergence of the optimal solution and improving load balancing degree greatly.