Engineering Evolutionary Algorithm to Solve Multi-objective OSPF Weight Setting Problem

Abstract

Setting weights for Open Shortest Path First (OSPF) routing protocol is an NP-hard problem. Optimizing these weights leads to less congestion in the network while utilizing link capacities efficiently. In this paper, Simulated Evolution (SimE), a non-deterministic iterative heuristic, is engineered to solve this problem. A cost function that depends on the utilization and the extra load caused by congested links in the network is used. A goodness measure which is a prerequisite of SimE is designed to solve this problem. The proposed SimE algorithm is compared with Simulated Annealing. Results show that SimE explores search space intelligently due to its goodness function feature and reaches near optimal solutions very quickly.