Search trajectory networks: A tool for analysing and visualising the behaviour of metaheuristics

A large number of metaheuristics inspired by natural and social phenomena have been proposed in the last few decades, each trying to be more powerful innovative than others. However, there is a lack accessible tools analyse, contrast visualise behaviour when solving optimisation problems. When metaphors are stripped away, these algorithms different their behaviour? To help answer this question, we propose data-driven, graph-based model, search trajectory networks (STNs) order directly types metaheuristics. One strength our approach that it does not require any additional sampling or algorithmic methods. Instead, models constructed from data gathered while We present methodology, consider detail two case studies covering both continuous combinatorial optimisation. In terms metaheuristics, cover main current paradigms: evolutionary, swarm, stochastic local approaches.