O. Berk Olmez

Problem definition. Services are increasingly offered through both physical and virtual channels by the same provider. We study a service provider’s decision about whether and how to pool servers across in-person (physical) and online (virtual) channels. We consider an environment inspired by urgent care, where customers make decisions based on observable expected delay times, and where customers’ channel preference is heterogeneous. Methodology/results. The usual goal of pooling servers is to improve utilization and reduce waiting. We use a queueing model to show that pooling resources between physical and virtual channels can lead to unprofitable demand shifts from one channel to the other under moderate to high offered loads, and this effect can paradoxically also increase expected waiting times. Managerial implications. A service provider may be better off maintaining dedicated servers for each channel, particularly when the profitability of the service varies across channels. This allows the service provider to capture the additional demand by adding a second channel without losing demand from the more-profitable channel. Some of the utilization benefits of pooling can also be achieved with dedicated servers by adding occasional flexibility for an idle server to offer service in the other channel, rather than full pooling. Alternatively, the service provider could pool servers across channels but use a probability-based priority service discipline, which allows finer control over customer flow under higher offered loads.

As the digital landscape continues to shape customer expectations, understanding the value customers place on virtual and in-person services has become increasingly important. This study leverages data from the undergraduate academic advisement operation at a public university in New York to explore the impact of service delivery channels on customer choices. By employing econometric models, we examine the influence of appointment attributes, such as delivery channel, date, time, duration, and distance, on customer preferences. The study addresses four key research questions: (1) How do customers value virtual appointments compared to in-person appointments? (2) Which appointment attributes have the most significant impact on customer decisions? (3) Are there demographic differences in the utilization of virtual and in-person services? (4) What factors contribute to appointment no-shows or cancellations? The findings of this research have implications for healthcare operations, as the popularity of telehealth and virtual appointments has grown.

Households produce large amounts of waste cooking oil (WCO), which should be disposed properly to avoid its negative impacts on the environment. Using WCO as raw material in biodiesel production is an effective disposal method. In this study, we focus on designing an efficient WCO collection network, in which households deposit their WCO into the bins to be placed at several collection centers (CCs), which are then regularly collected by capacitated vehicles. We study the problem of determining (i) the number and location of the CCs, (ii) the assignment of households to the CCs, respecting a certain coverage distance threshold, (iii) the number of bins to place at each CC, and (iv) the vehicle routes navigating between the biodiesel facility and CCs to minimize the sum of fixed cost of opening CCs and placing bins at each CC and routing cost. To solve the proposed location-routing problem, we develop a novel matheuristic based on Variable Neighborhood Search, which incorporates an integer programming model to determine the bin allocation and assignment decisions. We evaluate the solution performance of different variants of the proposed algorithm by performing an extensive computational study on a set of hypothetical and case instances. The results show that the proposed matheuristic provides superior solutions with respect to benchmark algorithms.