Self-adaptive evolutionary network coding algorithm: A constraint handling approach

This paper presents a self-adaptive evolutionary network coding algorithm (SA-ENCA) that minimizes the resources of network coding while achieving the target throughput of multicast. The idea is to adaptively engage infeasible solutions as well in searching for better solutions. This is achieved by assigning fitness to the infeasible solutions by balancing corresponding objective function values against constraint violations. Dealing with the constrained network coding problems in an unconstrained manner, SA-ENCA does not suffer from the drawbacks of existing approaches. In other words, it is able to effectively cope with selection noise and automatically discover a feasible seed in the course of evolution. Empirical study has adduced grounds for the effectiveness of the proposed approach.