Abstract
Omnichannel retailers are reinventing stores to meet the growing demand of the online channel. Several retailers now use stores as supporting distribution centers to offer quicker Buy-Online-Pickup-In-Store (BOPS) and Ship-From-Store (SFS) services. They resort to in-store picking to serve online orders using existing assets. However, in-store picking operations require picker carts traveling through store aisles, competing for store space, and possibly harming the offline customer experience. To learn picking policies that acknowledge interactions between pickers and offline customers, we formalize a new problem called Dynamic In-store Picker Routing Problem (diPRP). This problem considers a picker that tries to pick online orders (seeking) while minimizing customer encounters (hiding) - preserving the offline customer experience. We model the problem as a Markov Decision Process (MDP) and solve it using a hybrid solution approach comprising mathematical programming and reinforcement learning components. Computational experiments on synthetic instances suggest that the algorithm converges to efficient policies. We apply our solution approach in the context of a large European retailer to assess the proposed policies regarding the number of orders picked and customers encountered. The learned policies are also tested in six different retail settings, demonstrating the flexibility of the proposed approach. Our work suggests that retailers should be able to scale the in-store picking of online orders without jeopardizing the experience of offline customers. The policies learned using the proposed solution approach reduced the number of customer encounters by up to 50%, compared to policies solely focused on picking orders. Thus, to pursue omnichannel strategies that adequately trade-off operational efficiency and customer experience, retailers cannot rely on actual simplistic picking strategies, such as choosing the shortest possible route.

Abstract
In several online retail contexts, such as grocery retailing, customers have to be present at the moment of delivery, that is, an attended home delivery service is in place. This requirement adds new challenges to this channel, often leading to narrow profitability. From an operations perspective, this service is performed with the retailer offering multiple time slots for the customer to choose from. Retailers target a cost-efficient delivery process that also accounts for customers' preferences by properly managing the options to show to customers, that is, time slot management. This study analyzes a dynamic slotting problem, that is, choosing the best slots to show for each customer, which is close to many practical cases pursuing a customer service orientation. We study two new strategies to improve customer service while satisfying cost-efficiency goals: (i) enforcing a constraint on the minimum number or percentage of slots to show to customers and (ii) integrating multiple days when tackling this challenging problem. Our results show under which conditions these proposed strategies can lead to win-win situations for both customer service and profit.

Abstract
Following the widespread interest of both the scientific community and companies in using autonomous vehicles to perform deliveries, we propose the 'Drone-Assisted Vehicle Routing Problem with Robot Stations' (VRPD-RS), a problem that combines two concepts studied in the autonomous vehicles literature: truck-drone tandems and robot stations. We model the VRPD-RS as a mixed-integer linear program (MILP) for two different objectives, the makespan and operational costs, and analyze the impact of adding trucks, drones, and robots to the delivery fleet. Given the computational complexity of the problem, we propose a General Variable Neighborhood Search (GVNS) metaheuristic to solve more realistic instances within reasonable computational times. Results show that, for small instances of 10 customers, where the solver obtains optimal solutions for almost all cases, the GVNS presents solutions with gaps of 0.7% to the solver for the makespan objective and gaps of 0.0% for the operational costs variant. For instances of up to 50 customers, the GVNS presents improvements of 21.5% for the makespan objective and 8.0% for the operational costs variant. Furthermore, we compare the GVNS with a Simulated Annealing (SA) metaheuristic, showing that the GVNS outperforms the SA for the whole set of instances and in more efficient computational times. Accordingly, the results highlight that including an additional drone in a truck-drone tandem increases delivery speed alongside a reduction in operational costs. Moreover, robot stations proved to be a useful delivery element as they were activated in almost every studied scenario.

Abstract
Optimising operations in bulk cargo ports is of great relevance due to their major participation in international trade. In inbound operations, which are critical to meet due dates, the product typically arrives by train and must be transferred to the stockyard. This process requires several machines and is subject to frequent disruptions leading to uncertain processing times. This work focuses on the scheduling problem of unloading the wagons to the stockyard, approaching both the deterministic and the stochastic versions. For the deterministic problem, we compare three solution approaches: a Mixed Integer Programming model, a Constraint Programming model and a Greedy Randomised algorithm. The selection rule of the latter is evolved by Genetic Programming. The stochastic version is tackled by dispatching rules, also evolved via Genetic Programming. The proposed approaches are validated using real data from a leading company in the mining sector. Results show that the new heuristic presents similar results to the company's algorithm in a considerably shorter computational time. Moreover, we perform extensive computational experiments to validate the methods on a wide spectrum of randomly generated instances. Finally, as managing uncertainty is fundamental for the effectiveness of these operations, distinct strategies are compared, ranging from purely predictive to completely reactive scheduling. We conclude that re-scheduling with high frequency is the best approach to avoid performance deterioration under schedule disruptions, and using the evolved dispatching rules incur fewer deviations from the original schedule.

Abstract
The recent evidence demonstrating the importance of perishables in terms of store choice and shopping experience makes these products a very interesting topic in many different research areas. Nevertheless, the production planning research has not been paying the necessary attention to the complexities of production systems of such items. The evidence that consumers of perishable goods search for visual and other cues of freshness, such as the printed expiry dates, triggered the development of a multi-objective lot-sizing and scheduling model taking this relevant aspect into account by considering it explicitly as an objective function. A hybrid genetic algorithm based on NSGA-II was developed to allow the decision maker a true choice between different trade-offs from the Pareto front. Computational experiments were based on a case study, reported in the literature, concerning a diary company producing yogurt.

Abstract
Integrated production and distribution planning have received a lot of attention throughout the years and its economic advantages are well documented. However, for highly perishable products this integrated approach has to include, further than the economic aspects, the intangible value of freshness. We explore, through a multi-objective framework, the advantages of integrating these two intertwined planning problems at an operational level. We formulate models for the case where perishable goods have a fixed and a loose shelf-life (i.e. with and without a best-before-date). The results show that the economic benefits derived from using an integrated approach are much dependent on the freshness level of products delivered.