The "Last Mile Delivery" (LMD) refers to the last and most inefficient part of the supply chain. This is caused by the spatial distribution of disperse small receiving points, the ever-growing demand for faster shipment and the new time constraints of deliveries. Moreover, the small urban vehicles used for package distribution make LMD the most polluting part of the supply chain. This study describes the existing methods to improve the efficiency of LMD, its challenges and makes a conceptual proposal. The state-of-the-art techniques are categorised according to their data sources, the theoretical framework of the proposed models and the dynamic nature of the solutions.