Integrating chained P system with genetic rules for scheduling cloud two-stage tasks

This paper proposes a new method to optimize two-stage task scheduling problem in cloud environment based on the integrating chained P system with genetic rules (CPG system). Task scheduling problem is the main problem in Cloud Computing. The current scheduling optimization methods have some shortcomings, such as high complexity and easy fall into local optimization. The novel CPG system expands structure and rules expression and provides new evolutional model for traditional P system. The parallelism, uncertainty and elite retention mechanism of CPG system can solve the problem of fall into local optimization and accelerate the convergence speed of task scheduling in cloud environment. At the same time, in the decoding process of the task scheduling, knowledge-based method which inspired by Johnson’s rule is used to sort the task execution order for each Virtual Machine, so as to minimum delay time. Finally, simulations are conducted to verify CPG performs better than Johnson’s rule and Genetic algorithm (GA). The results demonstrate the validity of the CPG.