This research improves large-scale optimisation by combining problem decomposition with machine learning. By identifying similarities between subproblems, it predicts solutions instead of solving each independently, reducing computational cost. The approach enhances efficiency in logistics and extends to applications such as healthcare scheduling and transport network design.

This research seeks to reduce the energy consumption of 4G and 5G networks—currently about 3% of global usage—by identifying the factors that drive it. By modelling how elements like signal noise affect energy demands in antennas and processing hardware, the project aims to guide the design of more efficient, sustainable mobile networks.