Revenue Management deals with the question of when and at what price a product should be sold to which customer. The objective is to maximize revenue while making the best possible use of available resources. Forecasting and optimization models as well as simulations are used to determine solutions.

Revenue Management in Make-to-Order Production

Make-to-order production is one of the most recent application areas of Revenue Management and has therefore received relatively little research attention to date. However, production capacity is also perishable in make-to-order production and should be used optimally. Once production comes to a standstill, the unused production time cannot be utilized afterwards. This raises the question of which orders should be accepted and which should perhaps be rejected in order to maximize revenue; if too many low-value orders are accepted, later-arriving, higher-value orders may have to be rejected due to insufficient remaining production capacity (revenue displacement).

However, it is uncertain whether higher-value orders will actually arrive at a later point in time. In this case, rejecting too many low-value orders may result in production capacity not being fully utilized (revenue loss). Consequently, a balance between revenue displacement and revenue loss must be found.

In contrast to classical order acceptance models, Revenue Management does not evaluate incoming orders solely based on available production capacity when deciding whether to accept or reject them. Instead, demand forecasts and the value of the demand are also taken into account.

A particular characteristic of make-to-order production is that important, recurring customers, who consequently have a high customer value, may pay lower prices than new customers due to special conditions, for example. Consequently, different customer groups have different values in terms of both customer value and demand value.

Thus, the competing objectives of achieving target service levels for each customer group and maximizing contribution margins are important aspects that should be considered in research on Revenue Management in make-to-order production.

Revenue Management in the Cruise Industry

Cruises are becoming increasingly popular in all parts of the world. Accordingly, the number of passengers transported has grown and continues to grow steadily. At the same time, the cruise products offered are becoming more complex, while the cruise ships in operation are becoming larger and larger.

Similar to other industries, such as the airline and hotel industries and container transportation, the cruise industry can benefit from the use of Revenue Management methods and tools. The objective is to sell the available capacity within cruise networks as profitably as possible. This does not necessarily mean achieving the maximum possible utilization of the available capacity.

This research project investigates the potential for the successful application of Revenue Management in the cruise industry. To this end, mathematical optimization models for capacity allocation are developed, followed by booking acceptance control policies.

Revenue Management in Container Transportation

Which container slot on a ship should be sold to which customer, at what point in time, and at what price? This is the key question addressed by Revenue Management in container shipping.

Revenue Management systems have been successfully implemented in various industries, such as airlines, hotels, and car rental. The container shipping industry has similar characteristics and is therefore an interesting field of research for the application of these methods.

The limited capacity of the ship should be sold as efficiently and profitably as possible before departure. To achieve this, customers are segmented and booking limits are assigned to the different segments. Based on demand forecasts derived from historical and current data, mathematical models are used to determine these booking limits as effectively as possible.

To date, this industry has received relatively little attention in Revenue Management research and therefore offers considerable potential for new approaches.

The dissertation by Dr. Sebastian Zurheide, developed and published at the Institute of Quantitative Business Research and Business Information Systems:

"Zurheide, Sebastian: Revenue Management in Liner Shipping. Schriftenreihe Logistik-Management
in Forschung und Praxis, Volume 53, Verlag Dr. Kovač, Hamburg, 2015."

addresses this important research area. New approaches to Revenue Management in container shipping are presented and tested using extensive simulations.

The aim of this project is to model a holistic cyber-physical energy system. By integrating physical relationships, information and communication technologies (ICT), and market mechanisms, key challenges of future energy systems, such as sector coupling, decentralized generation, and high degrees of flexibility, can be taken into account. The focus is on optimizing the system with regard to the energy policy triangle, with the objectives of resilience, sustainability, and economic efficiency.

An important objective is to avoid greenhouse gas (GHG) emissions as quickly as possible in order to address the challenges of climate change. In Germany, most reductions have been achieved through the extensive expansion of renewable energies such as wind and photovoltaics.

Renewable energy is predominantly generated in the form of electricity. Therefore, ongoing GHG reductions are mainly taking place in the electricity sector. In order to decarbonize the entire energy system, electricity should also be used in other sectors, which leads to the concept of sector coupling.

During the energy transition, individual large-scale generators are being replaced by geographically distributed, supply-dependent decentralized energy resources (DER) with different energy qualities.

Since energy conversion, for example from electricity via power-to-gas to hydrogen or further to synthetic methane, is always associated with efficiency losses, direct electrification of these sectors is one part of this transformation. However, this conversion pathway creates opportunities for long-term energy storage to compensate for seasonal fluctuations in energy generation.

In order to maintain today's high level of supply reliability, new concepts for providing flexibility must be developed in this context.

Creating new flexibility options that can be provided both on the generation side and on the demand side requires specialized, highly developed information and communication technologies (ICT) to achieve the necessary high information density.

In addition, sector coupling takes place not only in the primary physical part of the system, but also in the secondary digital part.

However, the energy transition and sector coupling require changes not only on the physical and digital sides, but also on the market side. The detailed availability of data, for example through the ongoing rollout of smart meters, and the lack of tradability of small amounts of energy produced by DERs will lead to potential new market concepts.

One example is virtual power plants, which make use of geographically distributed generation and aggregation effects by combining different technologies, as well as improved forecasting possibilities resulting from stochastic effects when aggregating a large number of generators.

Despite the large number of existing energy models, these usually consider only a single sector. Once more sectors are included, the modeling depth must be reduced due to otherwise excessive complexity and increasing computational effort.

Although simplifications make it possible to simulate the overall system, this approach reduces the accuracy of the results. Market-related aspects are also often considered only for individual sectors.

The aim of this project is to develop a co-simulation platform that takes dynamic effects on the physical side into account and is capable of simulating event-based ICT models and protocols.

In addition, optimization models and algorithms are being developed and investigated that allow coupled effects to be captured within this holistic approach and optimal operating conditions to be derived while taking into account the energy policy triangle in Germany.

Further information about the project can be found on the project website .

Natural disasters as well as human-made disasters can have devastating consequences for the affected region and society. The objective of humanitarian logistics is therefore to minimize the negative effects of disasters through appropriate measures.

However, the work of humanitarian organizations is complicated by the enormous uncertainty regarding the location, timing, and scale of a disaster. In order to provide affected people with necessary relief supplies, such as water or medication, as quickly as possible, quantitative optimization models are used to support decision-making.

Development of an Efficient Procedure for Solving Two-Stage Stochastic Optimization Problems in Disaster Management

One of the most important measures for preparing for a natural disaster is the storage of relief supplies, such as medication and food, at selected locations.

This enables humanitarian organizations to provide assistance to affected people as quickly as possible after a disaster occurs. However, decisions regarding the location and type of storage are complicated by a high degree of uncertainty, as it is not known when, where, or to what extent a disaster will occur.

To support humanitarian organizations in their decision-making, an efficient procedure for solving realistic two-stage stochastic problems was developed as part of this dissertation. The procedure makes it possible to quickly initiate appropriate measures for the storage of relief supplies, even when disaster warnings are issued at short notice.

The dissertation by Dr. Emilia Grass, completed at the Institute of Quantitative Business Research and Business Information Systems:

“Grass, Emilia: Development of an Efficient Procedure for Solving Two-Stage Stochastic Optimization Problems in Disaster Management.
Dissertation, Hamburg University of Technology, 2017."

addresses this important research area.

Decision-Making in Humanitarian Logistics – A Multi-Criteria Transshipment Model for the Allocation of Relief Supplies in Ongoing Relief Operations

The dissertation deals with the planning of relief supply distribution in humanitarian operations while taking different objectives into account.

Quantitative planning models are used to enable optimized ordering, storage, and distribution of supplies.

Specific aspects considered in the modeling include, among others, a multi-period perspective and the consideration of uncertain future developments, such as the spread of an epidemic to additional regions.

A scenario-based analysis is intended to provide the respective decision-maker with the necessary information to make decisions about, for example, reordering relief supplies and their distribution.

The dissertation by Dr. Emilia Grass, completed at the Institute of Quantitative Business Research and Business Information Systems:

"Rottkemper, Beate: Decision-Making in Humanitarian Logistics – A Multi-Criteria Transshipment Model for the Allocation of Relief Supplies in Ongoing Relief Operations.
Dissertation, Hamburg University of Technology, 2016."

addresses this important research area.

When making decisions about future courses of action, complete information is generally not available. Such decisions must therefore be made under uncertainty.

Robust Optimization (RO) is a subfield of Operations Research (OR) that deals with proactive decision-making under uncertainty, i.e., it attempts to take uncertainty into account already during the planning process.

A resulting optimal course of action is considered robust if it remains feasible under all considered realizations of uncertainty.

The objective of this project is to investigate whether and how Business Analytics methods, such as machine learning (ML) and Operations Research (OR), can be used to reduce uncertainty and create robust plans for berth allocation at container terminals.

Optimization of Berth Allocation at Container Terminals Under Uncertainty

Using berth allocation as an example, this research investigates how the information required for decision-making under uncertainty can be derived from historical data.

The Berth Allocation Problem (BAP) aims to assign berths at the quay to arriving container ships.

A ship's berth is defined as a sufficiently large section of the quay that is reserved for the ship for a specified period of time. The berthing time marks the beginning of the ship's stay at the berth, which ends when the ship departs.

The berthing time must be at least as long as the handling time required for loading and unloading the containers.

Therefore, essential information includes, in particular, when container ships arrive at the container terminal and the volume of containers to be handled, i.e., how long the handling time will be.

These very factors are subject to considerable uncertainty in reality, making the Berth Allocation Problem an interesting subject for Robust Optimization.

Furthermore, the availability of data from the Automatic Identification System (AIS), which provides static and dynamic ship data at a high temporal resolution, provides a good basis for forecasting ship arrival times and berth occupancy times at container terminals using machine learning.

Berth allocation provides the basis for planning subsequent processes at the container terminal. Changes to the berth plan therefore also lead to changes in the planning of all subsequent steps.

Depending on the lead time, optimal or good solutions for subsequent processes may no longer be determined or implemented if short-term changes to the berth plan become necessary. In any case, replanning costs are incurred.

Therefore, in this context, berth allocation should not be understood merely as the optimization of a capacity bottleneck or utilization. It is also necessary to provide a stable basis for planning subsequent processes at the container terminal.