In this thesis, the aim is to propose more realistic OR models and to improve logistics decision-making by (1) considering stochastic input parameters, and (2) making use of historic data. For this purpose, we focus on two real-life problem settings. The first problem setting deals with a stochastic vehicle routing problem with time windows, based on a real-life case study. Customer-specific demand distributions are estimated from historic data using a mixture density network approach and by considering contextual information (e.g., packaging type). Time windows and a complex objective function minimizing both distance- and time-related costs are considered. A two-stage stochastic programming approach with recourse is used to model the problem, including expected violations of time windows, maximum route duration, maximum allowable driving time and other time-related constraints in the recourse cost function. A heuristic algorithm is used to solve the problem and results show that including uncertainty by means of historical data in the planning phase, results in reduced costs. In the second problem setting, a cooperative hospital supply chain with a central healthcare hub is compared with a traditional one, to investigate under which circumstances it is beneficial to move to a cooperative system. First, the concept of a cooperative supply chain is studied (advantages, disadvantages, characteristics etc.) using quantitative research methods (interviews and survey). Next, based on these insights, an inventory-routing problem is modelled, comparing a traditional system with a cooperative one, including real-life characteristics (e.g. uncertainty in demand, emergency deliveries, time windows). Results show that a cooperative hospital supply chain with a central healthcare hub may result in reduced costs while maintaining the same service level. At a higher service level, a higher inventory holding cost and a higher standard deviation of the demand, the cooperative scenario remains beneficial.