Despite the increased availability of data through sensors and other sources, real-world route planning often involve situations where customer demands are (partially) unknown during the route planning phase and only revealed upon arrival at the customers’ locations. In this dissertation, we propose and explore innovative and effective transportation solutions for two extensions to the vehicle routing problem with stochastic demands. In the first part of the thesis, we focus on situations where vehicle restrictions prohibit deliveries to customers with large, heavy, and polluting vehicles. To address the challenges posed by vehicle restrictions near customer locations, we consider a two-echelon distribution network. In this network, large trucks operate on the first echelon to leverage economies of scale, while smaller and environmentally friendly vehicles are employed on the second echelon to adhere to the vehicle restrictions imposed. This approach allows for efficient delivery operations while complying with the limitations in certain areas. We propose two efficient solution procedures based on column generation. Additionally, we reduce the number of customer combinations for which the chance constraint needs to be verified by imposing feasibility bounds on the stochastic customer demands. In the second part of the thesis, we shift to a single-echelon distribution system and allow partial deliveries. To ensure equitable allocation of resources among all customers, we impose that the minimum expected fill rate across all customers meets a predefined threshold. To take this into account, a priori routing and allocation decisions have to be taken simultaneously. However, the allocation decisions may be adjusted in response to observed demands. We develop a dynamic programming model to compute the minimum fill rate for a given route and develop a branch-price-and-cut algorithm to solve the corresponding vehicle routing problem.