Evolutionary algorithm and multifactorial evolutionary algorithm on clustered shortest-path tree problem

In literature, Clustered Shortest-Path Tree Problem (CluSPT) is an NP-hard problem. Previous studies focus on approximation algorithms which search for optimal solution in relatively large space. Thus, these consume a amount of computational resources while the quality obtained results lower than expected. order to enhance performance process, this paper proposes two different approaches are inspired by perspectives analyzing CluSPT. The first approach intuition narrow down space reducing original graph into multi-graph with fewer nodes maintaining ability find solution. problem then solved proposed evolutionary algorithm. This performs well those datasets having small number edges between clusters. However, increase size would cause excessive redundant that pressurize searching potential solutions. second overcomes limitation breaking set simple graphs. Every corresponding mutually exclusive From point view, could be modeled bi-level optimization includes nested spaces. Accordingly, Nested Local Search Evolutionary Algorithm (N-LSEA) introduced glscluspt, upper level uses algorithm Genetic Algorithm. Due neighboring characteristics local step level, reduced graphs share common traits among each others. Multi-tasking (MLSEA) take advantages underlying commonalities exploiting implicit transfer across similar tasks multi-tasking schemes. improvement experimental over N-LSEA via scheme inspires future works apply M-LSEA graph-based problems, especially optimization.