Fractional Political Optimizer-Based Switch Migration in Software-Defined WAN for Load Balancing with Deep Q Network
Bilal Ashfaq Ahmed, Norris Syed Abdullah, Mohd Foad Rohani, Saima Waseem · Cybernetics & Systems · 2022
In order to address the load imbalance problem, a successful technique known as Fractional Political Optimizer (FPO)-based switch migration is modeled. The suggested FPO is created by integrating the Political Optimizer (PO) with Fractional Calculus (FC). DQN receives the controller’s initial load value in order to generate the anticipated load. By calculating the load value based on CPU utilization, controller maximum capacity, and input packet volume, load imbalance issues are identified. Switch migration is carried out using FPO based on fitness measure, which takes into account migration factors, such as migration cost, projected load, controller capacity, and load balancing rate (LBR). To complete the switch migration procedure, the load on the controller is assessed using the data packets and a threshold value. For the NSFNet topology, the proposed FPO-based switch migration approach achieved the following results: 17.286 kbps, 0.1191 s, 48.797, −0.238, 0.1231 s, and 5.0391 kbps, respectively.