Multi-criteria Optimization Algorithms specially tailored for Programmable Networks with Quality Management

The new generation of programmable networks should allow efficient bandwidth control and quality of service (QoS) management of different delay- and loss-sensitive services. To this end, we propose specially tailored evolutionary algorithms for solving routing problems in next-generation networks such as SDN, IoT, and 5/6G. In particular, we optimize network routing for three critical parameters: throughput, loss, and delay, which forms a multi-criteria optimization problem. We propose two evolutionary algorithms: NSGA-II and ISPEA2 to solve the problem and perform simulation tests for selected real and artificial network topologies of various sizes. The results obtained show that NSGA-II performs significantly better than ISPEA2.