Scheduling Strategy for Batch Scientific Workflows based on Improved Discrete Binary Particle Swarm Optimization in Cloud Platform

The batch scientific workflow scheduling strategy based on discrete binary group optimization (DB-S) is a mechanism that uses the power of group intelligence to solve multi-objective optimization problems. DB-S algorithm was originally proposed by Dorigo et al. in 2003, and has recently been extended to support different types of discrete objects, such as particles and molecules, also known as discrete binary particle swarm optimization (DBPSO). This is a new method that combines the advantages of classical evolutionary algorithms with the advantages of traditional optimization methods. Aiming at the security threats faced by the cloud workflow scheduling problem, firstly, the cloud model is used to quantify the security of tasks and virtual machine resources, and the security cloud similarity is used to measure users' satisfaction with the security of virtual machine resources allocated by tasks; Then, a cloud workflow scheduling model considering security, completion time and usage cost is established, and a cloud workflow scheduling algorithm based on discrete particle swarm optimization is proposed; Finally, the proposed algorithm is simulated. The experimental results show that compared with similar algorithms, this algorithm has better performance in terms of security utility value, completion time, usage cost and load balancing deviation.