Complex Network Analysis of Evolutionary Algorithms Applied to Combinatorial Optimisation Problem
This research analyses the development of a complex network in an evolutionary algorithm (EA). The main aim is to evaluate if a complex network is generated in an EA, and how the population can be evaluated when the objective is to optimise an NP-hard combinatorial optimisation problem. The population is evaluated as a complex network over a number of generations, and different attributes such as adjacency graph, minimal cut, degree centrality, closeness centrality, betweenness centrality, k-Clique, k-Club, k-Clan and community graph plots are analysed. From the results, it can be concluded that an EA population does behave like a complex network, and therefore can be analysed as such, in order to obtain information about population development.
KeywordsEvolutionary algorithm complex network flow shop scheduling
Unable to display preview. Download preview PDF.
- 2.Zelinka, I., Davendra, D., Enkek, R., Jasek, R.: Do Evolutionary Algorithm Dynamics Create Complex Network Structures? Complex Systems 20, 127–140, 2, 0891-2513Google Scholar
- 4.Davendra, D., Onwubolu, G.: Forward backward transformation. In: Onwubolu, G., Davendra, D. (eds.) Differential Evolution. SCI, vol. 175, pp. 35–80. Springer, Heidelberg (2009)Google Scholar
- 8.Wolfram Website, http://www.wolfram.com/