Fitness-distance-ratio particle swarm optimization
Christopher W. Cleghorn, Andries Petrus Engelbrecht · Proceedings of the Genetic and Evolutionary Computation Conference · 2017
At present the fitness-distance-ratio particle swarm optimizer (FDR-PSO) has undergone no form of theoretical stability analysis. This paper theoretically derives the conditions necessary for order-1 and order-2 stability under the well known stagnation assumption. Since it has been shown that particle stability has a meaningful impact on PSO's performance, it is important for PSO practitioners to know the actual criteria for particle stability. This paper validates its theoretical findings against an assumption free FDR-PSO algorithm. This empirical validation is necessary for a truly accurate representation of FDR-PSO's stability criteria.