The Vehicle Routing Problem with Simultaneous Pickup and Delivery and Time Window (VRPSPDTW) is a challenging optimisation problem in logistics and distribution management, which has received increasing attention from researchers. Most of the existing algorithms, which are able to search for high-quality solutions under advance preference settings, are unable to find multiple high-quality solutions in a single run. Therefore, this thesis innovatively suggests that the Constrained Multi-Objective Evolutionary Algorithm, CMO-VRP, is utilised to solve VRPSPDTW. A two-stage search strategy is utilised, where constraints are ignored and only convergence and diversity are considered in the first stage, and multiple swarms are utilised in the second stage, where constraints, convergence and diversity are considered simultaneously. The algorithm in this thesis is shown to be highly competitive with current algorithms through an experimental comparison of 65 algorithms.