This thesis investigates the optimization of Same-Day Delivery (SDD) in the context of a Dynamic and Stochastic Vehicle Routing Problem with Time-Windows (DSVRPTW). A central focus is the concept of pre-releasing; the process of assigning an order to a specific route and preparing it for delivery, effectively fixing theorder to the route. This happens before all orders are known, restricting the possibleroutes and thus increasing the costs.

By conducting computational experiments using real-life data from a European e-grocery, the study evaluates various pre-releasing heuristics. The most effective heuristic, which delays pre-releasing until the last possible moment before the route departs, results in a reduction of optimization costs of 20% compared to current operations. When allowing more than 2 time-windows per truck, the reduction increases to 32%.

The research also addresses the critical practical constraint of pre-releasing capacity, which represents the maximum number of orders that can be pre-released within an hour. Simulation analysis reveals that using the last-minute heuristic only slightly increases the pre-releasing rate compared to the current operations of companies. To address this constraint efficiently, the most cost-effective solution is to invest in additional warehouse workforce. Alternatively, pre-releasing orders one hour early incurs a 1% increase in costs, while pre-releasing two hours early results in a 4% cost increase. Furthermore, initiating pre-releasing activities in the morning rather than the previous night can lead to savings of up to 2%.

The e-grocery expresses a preference for a constant pre-releasing rate equal to the pre-releasing capacity. This thesis proposes eight additional heuristics that determine which orders should be pre-released in addition to those identified by the last-minute heuristic. While the results from limited data were inconclusive, on average, the best heuristics incur an 8% higher cost compared to only pre-releasing at the last moment by pre-releasing orders based on proximity to or distance from the warehouse. Notably, pre-releasing orders with closer time-windows appeared to be preferable. Implementing this strategy would allow the pre-release capacity to be reduced from 200 to 150, resulting in a savings of five full-time pickers in the warehouse.

Future research opportunities include testing the methods on different cases and datasets and estimating the probability of exceeding the pre-releasing capacity, which could be used for deciding whether or not it is necessary to pre-release extra orders.

Money laundering is an increasing problem for the global economy. To combat money laundering, banks use transaction monitoring models with particular thresholds to detect unusual transaction behaviour. However, it is a challenge to determine and evaluate the suitability of a threshold level to ensure that the risk of misclassification of transactions falls within the bank’s risk appetite. In the threshold tuning process, the suitability of a threshold level can be evaluated with a sample of the transactions below or above a threshold level which are reviewed by an analyst. One problem is that the review process of transactions during the threshold tuning process is timeconsuming. In addition, banks want to be able to quantify the risk of misclassification of transactions to determine whether this falls within their risk appetite. This underlines the need to develop a threshold tuning strategy to accelerate the threshold tuning process in which the risk of misclassification of transactions can be quantified to determine whether it falls within the bank’s risk appetite. To accelerate the threshold tuning process, a framework was developed and five threshold tuning strategies were established which evaluate the suitability of different threshold levels with a given strategy. In addition, several methods to determine a confidence interval were examined to quantify the risk of misclassification and to ensure that it falls within the bank’s risk appetite. The threshold tuning strategies were compared and evaluated on the required amount of reviews of transactions and the difference between the found and true threshold level using synthetic data sets. Overall, the bisection threshold tuning strategy is recommended, since this strategy resulted in the lowest number of required reviews of transactions and resulted in a small difference between the found and true threshold level. The results of the synthetic data sets were promising, but more experiments with preferably real transaction data or other distributions are required to further evolve and fully validate the framework and proposed bisection strategy. The work presented in this thesis contributed to a more risk-based approach to enhance the efficiency and effectiveness of the threshold tuning process of transaction monitoring models.