Using Swarm Intelligence to Optimize the Energy Consumption for Distributed Systems

  •  Neil Bergmann    
  •  Yuk Chung    
  •  Xiangrui Yang    
  •  Zhe Chen    
  •  Wei Yeh    
  •  Xiangjian He    
  •  Raja Jurdak    


Large, distributed, network-based computing systems (also known as Cloud Computing) have recently gained significant interest. We expect significantly more applications or web services will be relying on network-based servers, therefore reducing the energy consumption of these systems would be beneficial for companies to save their budgets on running their machines as well as cooling down their infrastructures. Dynamic Voltage Scaling can save significant energy for these systems, but it faces the challenge of efficient and balanced parallelization of tasks in order to maximize energy savings while maintaining desired performance levels. This paper proposes our Simplified Swarm Optimization (SSO) method to reduce the energy consumption for distributed systems with Dynamic Voltage Scaling. The results of SSO have been compared to the most popular evolutionary Particle Swarm Optimization (PSO) algorithm and have shown to be more efficient and effective, reducing both the execution time for scheduling and makespan.

This work is licensed under a Creative Commons Attribution 4.0 License.