A Self-Adaptive Detoxify Particle Swarm Optimization Algorithm for Optimal Route Detection in Software Defined Wide Area Network

The significant reliance on data in enterprise networks requires efficient routing of user data. Traditional WANs use routing techniques like MPLS and DSL, but they only provide data forwarding services and limit network scalability. SDN addresses these issues by decoupling control and data planes, offering network automation, scalability, and policy-driven management. The adoption of gradient-based ML algorithms in SDN improves data routing. However, the complexity of deployment and potential drawbacks of gradient-based algorithms led to the development of less complex, more adaptive gradient-free Nature Inspired Algorithms (NIAs). Though NIAs like PSO and ACO are easier to implement, they have limitations such as premature convergence and increased processing delay. Hence, this research proposes a Self-Adaptive Particle Swarm Optimization (SAPSO) algorithm for optimal route detection in a scalable SDWAN. This approach tracks convergence, uses a flow entry algorithm, and encourages global search with a detoxify swarm function to find the optimal route in a scalable space.