Interaction networks have emerged as a valuable tool for analysing the behaviour of population-based metaheuristics from the swarm and evolutionary families. This approach represents algorithms as dynamic networks, where nodes correspond to individuals within the population, and edges depict the interactions between them. In this paper, we apply interaction networks to model and study Genetic Algorithms, a well-known representative of evolutionary algorithms. The focus is on examining the effects of different crossover operators on the interaction networks, as these are key mechanisms for information exchange within the population. Using real-world optimisation problems, the methodology was tested, and correlations between the algorithm’s performance and the network structures were identified. The results indicate that this framework offers a new perspective on how individual interactions influence overall algorithm behaviour. Additionally, the potential of interaction networks to compare and evaluate different algorithm variants is highlighted.