Optimization of urban transport is an ever-growing research area, especially with the massive success of e-commerce, the substantial demographic increase in urban areas, but also exploring interesting trends such as flexible multimodal itineraries. This work presents an innovative multi-stage approach where optimization methods such as greedy algorithms are adapted to solve a bi-objective vehicle routing problem with pickup and deliveries in urban areas with transfers. We highlight the performance of our approach on several instances with various sizes and characteristics.