Clemens van Dinther

Cluster Analysis of Smart Metering Data

The introduction of smart meter technology is a great challenge for the German energy industry. It requires not only large investments in the communication and metering infrastructure, but also a redesign of traditional business processes. The newly incurring costs cannot be fully passed on to the end customers. One option to counterbalance these expenses is to exploit the newly generated smart metering data for the creation of new services and improved processes. For instance, performing a cluster analysis of smart metering data focused on the customers’ time-based consumption behavior allows for a detailed customer segmentation. In the article we present a cluster analysis performed on real-world consumption data from a smart meter project conducted by a German regional utilities company. We show how to integrate a cluster analysis approach into a business intelligence environment and evaluate this artifact as defined by design science. We discuss the results of the cluster analysis and highlight options to apply them to segment-specific tariff design.

Sensitivity analysis of the economic benefits from electricity storage at the end consumer level

The article presents the results of simulations based on a linear optimization model of a storage system that calculates the economic benefits of distributed storage devices at the end consumer level by determining the cost optimal charge-discharge-schedule. The primary objective of the storage application is arbitrage accommodation. Particularly, parameters for a li-ion-based and a lead-acid-based storage system are simulated. All parameters of the model are varied and analyzed regarding their impact on the economic benefits. The simulation results quantify these impacts and show that the costs per storage capacity unit (EUR/kWh) and the efficiency degrees of the storage system have the highest impact. Additionally, the price spreads as well as the distribution of the market price curve in relation with the distribution of the consumers load curve (demand) influence the achievable benefits significantly. Overall, the model reveals a saving potential of 17% on total cost for the reference case.

A multidimensional procurement auction for trading composite services

Recently, static value chains have gradually been giving way to highly agile service value networks. This involves novel economic and organizational challenges. Added value for customers is created by feasible compositions of distributed service components. This work focuses on the design of a multidimensional procurement auction for trading service compositions and the analysis of strategies for service providers that participate in the procurement process. The mechanism implementation is incentive-compatible, so that it results in an equilibrium in which revealing the true multidimensional type (quality of service and valuation) is a weakly-dominant strategy for all service providers. Due to combinatorial restrictions imposed by the underlying graph topology, the winner determination problem can be solved in polynomial time, in contrast to computationally-intractable combinatorial auctions which cannot be solved this way. Furthermore, we provide a simulation-based analysis based on a reinforcement learning model of bundling and unbundling strategies of service providers that participate in the auction. Based on our results we discuss strategic recommendations for service providers depending on how they are situated within the network.

How to Coordinate Value Generation in Service Networks

The fundamental paradigm shift from traditional value chains to agile service value networks implies new economic and organizational challenges. As coordination mechanisms, auctions have proven to perform quite well in situations where intangible and heterogeneous goods are traded. Nevertheless, traditional approaches in the area of multidimensional combinatorial auctions are not quite suitable to enable the trade of composite services. A flawless service execution and therefore the requester’s valuation highly depends on the accurate sequence of the functional parts of the composition, meaning that in contrary to service bundles, composite services only generate value through a valid order of their components. The authors present an abstract model as a formalization of service value networks. The model comprehends a graph-based mechanism implementation to allocate multidimensional service offers within the network, to impose penalties for non-performance and to determine prices for complex services. The mechanism and the bidding language support various types of QoS attributes and their (semantic) aggregation. It is analytically shown that this variant is incentive compatible with respect to all dimensions of the service offer (quality and price). Based on these results, the authors numerically analyze strategic behavior of participating service providers regarding possible collusion strategies.

Benchmarking electric vehicle charging control strategies

Electric vehicles (EVs) are expected to become an important part of individual mobility. In order to reduce CO2 emissions and release the full potential for sustainable mobility, EVs need to be charged with energy from renewable energy sources (RES). We employ a deterministic linear optimization approach with different coordination objectives for each simulation scenario. The objectives are to minimize the individual average charging costs, maximize the average use of wind power or minimize the average load factor for the charging times of each EV customer. Customers have real life driving profiles from the German mobility panel and are distinguished in employees and retired with their respective driving behavior. We find that the wind power share used for charging can be nearly doubled for both groups under the respective strategy. Average costs are increased in comparison to the cost oriented strategy but are considerably lower as in the uncoordinated charging case.

Service Value Networks: Unleashing the Combinatorial Power of Service Mashups

The long tail phenomenon has been heavily discussed in recent years. What has been neglected so far is its striking relevance for electronic services. Whereas consumers expectations about information goods are often vague and transient, their requirements are pronounced and specific when it comes to the functional and non-functional characteristics of electronic services. Moreover, modular services can be combined and configured into service mashups that have the potential to meet complex consumer requirements. In this vein, the long tail phenomenon can be leveraged into a new dimension – the long valley, where every service exerts positive network externalities on the remaining services, thereby spurring an increase in supply and demand. The combinatorics of constructible service mashups are enabled by universally accessible service orchestration platforms known as Service Value Networks (SVNs). This article shall not only pave the way for a rising research area on a new business trend, but will also help business actors to harness the opportunities opened up by SVNs.

Strategic Learning by e-Service Suppliers in Service Value Networks

This article examines strategic behavior of e-Service suppliers that offer electronic services in complex Service Value Networks (SVNs). In SVNs, consumers request a bundle of e-Services, and the SVN acts as an aggregator of single service instances, automatically configuring services from different e-Service suppliers into a complex service bundle, which is then offered to the consumers. In this context, e-Service suppliers who want to maximize their business success need to configure their services according to the (for them unknown) preferences of the consumers. Current literature, however, does not offer much guidance on how to find a fitting service offer in this situation, especially when the suppliers are faced with a changing consumer base. For this reason, we study two specific learning regimes that are able to capture and deal with the inherent complexity of the corresponding strategy space. Besides finding beneficial service offers, e-Service suppliers might also learn collusive behavior if it aligns with individual incentives. The potential occurrence of such behavior is the second aim of our work. Our results show that even with relatively simple learning regimes, e-Service suppliers are able to find beneficial offers and learn to collude tacitly (and presumably legally), which increases their profits.

Clusteranalyse von Smart-Meter-Daten

ZusammenfassungDie Einführung der Smart-Meter-Technologie stellt die Energiewirtschaft in Deutschland vor große Herausforderungen. Neben hohen Investitionen in die Zähler- und Kommunikationsinfrastruktur ist auch die Neugestaltung vieler Geschäftsprozesse erforderlich. Da die neu entstehenden Kosten nur begrenzt an Endkunden übertragbar sind, gilt es die Aufwendungen der Energiewirtschaft durch neue Dienste und verbesserte Prozesse auf Basis von Smart Metering zu kompensieren. So ist durch die Clusteranalyse der detaillierteren Verbrauchsdaten eine deutlich feinere Kundensegmentierung auf Basis des zeitlichen Verbrauchsverhaltens möglich. Im Rahmen eines Smart-Metering-Projektes bei einem regionalen Energieversorger wurde eine Clusteranalyse für die real vorliegenden Kundenverbrauchsdaten entwickelt und in eine Business-Intelligence-Umgebung integriert. In diesem Beitrag beschreiben und evaluieren wir dieses Artefakt im Sinne der Design Science. Wir gehen dabei insbesondere auf die Ergebnisse der Clusteranalyse von Realdaten und den möglichen Einsatz zur segmentspezifischen Tarifgestaltung ein.AbstractThe introduction of smart meter technology is a great challenge for the German energy industry. It requires not only large investments in the communication and metering infrastructure, but also a redesign of traditional business processes. The newly incurring costs cannot be fully passed on to the end customers. One option to counterbalance these expenses is to exploit the newly generated smart metering data for the creation of new services and improved processes. For instance, performing a cluster analysis of smart metering data focused on the customers’ time-based consumption behavior allows for a detailed customer segmentation. In the article we present a cluster analysis performed on real-world consumption data from a smart meter project conducted by a German regional utilities company. We show how to integrate a cluster analysis approach into a business intelligence environment and evaluate this artifact as defined by design science. We discuss the results of the cluster analysis and highlight options to apply them to segment-specific tariff design.

Agent-based Simulation for Research in Economics

Financial theory and financial markets are complex constructs that are not always easy to understand. It is even not possible to explain the function of financial markets on basis of theoretical analysis, since theory normally assumes fully rational agents. [37] have presented the so called “No trade theorem” stating that in a state of efficient equilibrium there will be no trade on markets as long as all participants are fully rational (no noise traders or other uninformed traders) and the structure of information acqusition is common knowledge. It is obvious that the assumptions do not hold in real markets, since we observe asymmetric informed participants or technical traders who base their decisions on the analysis of market movements (and not on the assessment of the underlying value of a stock).

Designing Auctions for Coordination in Service Networks

The evolving service ecologies show new ways of value co-creation through combinations of multiple service components which are described in service offerings. An open issue in such a large service ecology is how to efficiently coordinate and price service offerings. Service offerings provide different functionality and quality. Customers need to distinguish their preferences on different combinations of service attributes. In this chapter we address this issue of service offerings allocation and introduce a structure design approach, Market Engineering, as an appropriate method to design such mechanism. In order to apply this approach to service systems we introduce a formal model and a definition of service value networks. Examples exemplify our approach and we show one possible step towards implementing such a mechanism.