Considering spatiotemporal evolutionary information in dynamic multi‐objective optimisation

Preserving population diversity and providing knowledge, which are two core tasks in the dynamic multi-objective optimisation (DMO), challenging since sampling space is time- space-varying. Therefore, spatiotemporal property of evolutionary information needs to be considered DMO. In present study, a sliding-time-window-based clustering method (SPC) proposed effectively solve problems (DMOPs). SPC, knowledge provided by saving historical data temporal dimension, spectral used divide saved into multiple neighbourhood subspaces for preserving spatial dimension. The SPC incorporated RM-MEDA compared with other recently state-of-the-art algorithms (DMOEAs) on 14 DMOPs introduced IEEE CEC2018. Simulation results demonstrate that capable enhancing tracking performance dynamically changing environments. Additionally, utilised an actual translation control problem immersed tunnel element. Results show outperforms knee point-based transfer learning terms both computational cost performance.