A Scalable Forecasting Framework to Predict COVID-19 Hospital Bed Occupancy

We present a scalable forecasting framework with a Monte Carlo simulation to forecast the short-term bed occupancy of patients with confirmed and suspected COVID-19 in intensive care units and regular wards. Our forecasts were a central part of the official weekly reports of the Bavarian State Ministry of Health and Care from May 2020 to March 2021.

Galactic Air Improves Ancillary Revenues with Dynamic Personalized Pricing

Ancillaries are a rapidly growing source of revenue for airlines, yet their prices are currently statically determined using rules of thumb and are matched only to the average customer or to customer groups. Offering ancillaries at dynamic and personalized prices based on flight characteristics and customer needs could greatly improve airline revenue and customer satisfaction. Through a start-up (Deepair) that builds and deploys novel machine learning techniques to introduce such dynamically priced ancillaries to airlines, we partnered with a major European airline, Galactic Air (pseudonym), to build models and algorithms for improved pricing. These algorithms recommend dynamic personalized ancillary prices for a stream of features (called context) relating to each shopping session. Our recommended prices are restricted to be lower than the human-curated prices for each customer group. We designed and compared multiple machine learning models and deployed the best-performing ones live on the airline’s booking system in an online A/B testing framework. Over a six-month live implementation period, our dynamic pricing system increased the ancillary revenue per offer by 25% and conversion rate by 15% compared with the industry standard of human-curated rule-based prices.

A Digital Twin for Decision Making on Livestock Feeding

Looking for an accurate and cost-effective solution to measure feed inventories, forecast the feed demand and allow feed suppliers to optimize inventories, production batches, and delivery routes.

Filling a Theater During the COVID-19 Pandemic

In the summer of 2020, Music Building Eindhoven (MBE) had to deal with the economic consequences of the COVID-19 pandemic for theater halls because governmental regulations were having a severe impact on the occupancy. In particular, MBE faced the challenge of determining how to maximize the number of guests in a theater hall while respecting social distancing rules. We have developed and implemented an optimization model based on trapezoid packings to address this challenge. The model showed that up to 40% of the normal capacity can be realized for a single show setting and up to 70% in cases where artists opt for two consecutive performances per evening without reusing seats. The solution was adopted by MBE with significant monetary and managerial benefits.

Practice Summary: Solving the External Candidates Exam Schedule in Norway

We developed a mixed-integer linear programming model to plan exam sessions for external candidates in the Vestfold region, Norway. With our model, the administration planned the last session of 2018, the two sessions of 2019, and the first session of 2020. The plans produced are of high quality and saved three weeks of person effort per session.

Inventory Management Using a Weekly Review (s, S) Policy at the Bank of Canada

While the inventory management problem faced by central banks is complicated in that they must deal with two-way shipments, complex costing agreements and insurance limits on inventory, we were able to develop an adapted version of an (s,S) policy that reduced the costs of the central bank’s inventory management by 25%. Implementation required very little change to the regular practices of the Bank and resulted in very few unintended consequences.

An Air Force Pilot Training Recommendation System Using Advanced Analytical Methods

The planning of individualized pilot training programs is an intensive process. Over 120 maneuvers are introduced into the training program over time while ensuring maneuver competencies. This work introduces a novel, deep-learning based approach for automatically generating training plans for pilot trainees to significantly reduce instructor pilot planning requirements.

Practice Summary: A Multiobjective Assignment Model for Optimal Socially Distanced Classrooms for the Spears School of Business at Oklahoma State University

In response to Oklahoma State University’s goal to provide safe face-to-face course options for as many students as possible in fall 2020, the Spears School of Business leadership developed a strategy based on data analytics and multiobjective optimization that allowed the implementation of a smooth transition plan for classes in fall 2020. OpenSolver was used (within Excel) to create an effective and efficient tool that successfully implemented classroom assignments that maximized face-to-face offerings while minimizing disruptions to the schedule, faculty, and registrar’s office.

A Decision Support System for Brewery Production Planning at Feldschlösschen

This article presents a decision support system (DSS) for multilevel production planning implemented at a Swiss brewery. The DSS supports supply chain executives in tactical and strategic decision making. The system consists of a user interface and visualization tool embedded in a cloud-based optimization framework.

Scheduling the Valley Baseball League

The Valley Baseball League is a collegiate summer baseball league with 11 teams playing 42 games between June and August. Producing a schedule has proved challenging for league management, because of the odd number of teams as well as a large number of league mandates and team-specific requests. As an amateur league with limited financial resources, there is pressure to best leverage the league schedule to maximize revenue for each team. A schedule generated by mathematical programming was enthusiastically adopted for the 2020 season and significantly outperformed previous manually produced schedules.