[Show abstract][Hide abstract] ABSTRACT: In this paper we demonstrate the potential of data analytics methods for location-based services. We develop a support system that enables user-based relocation of vehicles in free-floating carsharing models. In these businesses, customers can rent and leave cars anywhere within a predefined operational area. However, due to this flexibility, free-floating carsharing is prone to supply and demand imbalance. The support system detects imbalances by analyzing patterns in vehicle idle times. Alternative rental destinations are proposed to customers in exchange for a discount. Using data on 250,000 rentals in the city of Vancouver, we evaluate the relocation system through a simulation. The results show that our approach decreases the average vehicle idle time by up to 16 percent, suggesting a more balanced state of supply and demand. Employing the system results in a higher degree of vehicle utilization and leads to a substantial increase of profits for providers.
International Conference on Information Systems 2015, Fort Worth, TX; 12/2015
[Show abstract][Hide abstract] ABSTRACT: The rapid and ongoing evolution of mobile devices allows for increasing ubiquity of online handhelds, yet boosting the recent growth of social platforms. This development facilitates participation in social media for an enormous amount of individuals independently from time and location. When navigating through a city and especially when following activities worthy to be shared with others, people uncover their traces in both geographical and temporal dimension. Using these traces to spot popular areas in a metropolitan region is valuable to a broad variety of applications, reaching from city planning to venue recommendation and investment. We propose a density-based method to determine the attractiveness of areas based solely on spatial and contentual characteristics of Twitter activity. Furthermore, we show the relation of attached images, videos, or linked places to the activity users are engaged in and assess the explanatory power of Twitter messages in a geographical context.
2015 Business Analytics Congress, Fort Worth, TX; 12/2015
[Show abstract][Hide abstract] ABSTRACT: Flash crashes, perceived as sharp drops in market prices that rebound shortly after, have turned the public eye toward the vulnerability of information technology-based stock trading. In this paper, we explain flash crashes as the result of actions made by rational agents. We argue that the advancement of information technology (IT), which has long been associated with competitive advantages, may cause ambiguities with respect to the game form that give rise to a hypergame. We employ hypergame theory to demonstrate that a market crash constitutes an equilibrium state if players misperceive the true game. Once the ambiguity is resolved, prices readjust to the appropriate level, creating the characteristic flash-crash effect. By analyzing the interaction with herd behavior, we find that flash crashes may be an unavoidable systemic problem of modern financial markets.
Journal of Management Information Systems 04/2015; 31(4):88-108. DOI:10.1080/07421222.2014.1001258 · 2.06 Impact Factor
[Show abstract][Hide abstract] ABSTRACT: As a rapidly expanding market, carsharing presents a possible remedy for traffic congestion in urban centers. Especially free-floating carsharing, which allows customers to leave their car anywhere within the operator’s business area, provides users with flexibility, and complements public transportation. We present a novel method that provides strategic and operational decision support to companies maneuvering this competitive and constantly changing market environment. Using an extensive set of customer data in a zero-inflated regression model, we explain spatial variation in carsharing activity through the proximity of particular points of interests, such as movie theaters and airports. As an application case, as well as a validation of the model, we use the resulting indicators to predict the number of rentals before an expansion of the business area and compare it to the actual demand post-expansion. We find that our approach correctly identifies areas with a high carsharing activity and can be easily adapted to other cities.
[Show abstract][Hide abstract] ABSTRACT: Cloud computing promises the flexible delivery of computing services in a pay-as-you-go manner. It allows customers to easily scale their infrastructure and save on the overall cost of operation. However Cloud service offerings can only thrive if customers are satisfied with service performance. Allow-ing instantaneous access and flexible scaling while maintaining the service levels and offering competitive prices poses a significant challenge to Cloud Computing providers. Furthermore services will remain available in the long run only if this business generates a stable revenue stream. To address these challenges we introduce novel policy-based service admission control mod-els that aim at maximizing the revenue of Cloud providers while taking in-formational uncertainty regarding resource requirements into account. Our evaluation shows that policy-based approaches statistically significantly out-perform first come first serve approaches, which are still state of the art. Furthermore the results give insights in how and to what extent uncertainty has a negative impact on revenue.
European Journal of Operational Research 01/2015; forthcoming(2). DOI:10.1016/j.ejor.2015.01.027 · 2.36 Impact Factor
[Show abstract][Hide abstract] ABSTRACT: Successful companies are increasingly those companies that excel in the task of extracting knowledge from data. Tapping the source of ‘Big Data’ requires powerful algorithms combined with a strong understanding of the data used. One of the key challenges in predictive analytics is the identification of relevant factors that may explain the variables of interest. In this paper, we present a case study in predictive analytics in which we focus on the selection of relevant exogenous variables. More specifically, we attempt to predict the German electricity spot prices with reference to historical prices and a deep set of weather variables. In order to choose the relevant weather stations, we use the least absolute shrinkage selection operation (LASSO) and random forests to implicitly execute a variable selection. Overall, in our case study of German weather data, we manage to improve forecasting accuracy by up to 16.9% in terms of mean average error.
Journal of Decision System 01/2015; 24(1). DOI:10.1080/12460125.2015.994290
[Show abstract][Hide abstract] ABSTRACT: Free-floating carsharing is a young and rapidly expanding market that allows customers to end their rentals anywhere within the business area of the provider. Through this flexibility it complements public transportation and reduces the environmental footprint of the transportation sector. We present a novel data analytics methodology that supports companies – from local start-ups to global players – in maneuvering this constantly growing and changing market environment. Using a large set of rental data, we derive indicators for the attractiveness of certain areas based on points of interest in their vicinity, such as shopping malls, movie theaters, or train stations. In a case study of Berlin we use these indicators to accurately identify promising regions for an expansion of the business area. However, the methodology introduced in this paper can also improve operational decisions of the service provider and is applicable to a wide range of other location-based services.
International Conference on Information Systems 2014, Auckland; 12/2014
[Show abstract][Hide abstract] ABSTRACT: Natural disasters, such as earthquakes, tsunamis and hurricanes, cause tremendous harm each year. In order to reduce casualties and economic losses during the response phase, rescue units must be allocated and scheduled efficiently. As this problem is one of the key issues in emergency response and has been addressed only rarely in literature, this paper develops a corresponding decision support model that minimizes the sum of completion times of incidents weighted by their severity. The presented problem is a generalization of the parallel-machine scheduling problem with unrelated machines, non-batch sequence-dependent setup times and a weighted sum of completion times – thus, it is NP-hard. Using literature on scheduling and routing, we propose and computationally compare several heuristics, including a Monte Carlo-based heuristic, the joint application of 8 construction heuristics and 5 improvement heuristics, and GRASP metaheuristics. Our results show that problem instances (with up to 40 incidents and 40 rescue units) can be solved in less than a second, with results being at most 10.9% up to 33.9% higher than optimal values. Compared to current best practice solutions, the overall harm can be reduced by up to 81.8%.
European Journal of Operational Research 06/2014; 235(3):697–708. DOI:10.1016/j.ejor.2013.10.029 · 2.36 Impact Factor
[Show abstract][Hide abstract] ABSTRACT: Due to the integration of intermittent resources of power generation such as wind and solar, the amount of supplied electricity will exhibit unprecedented fluctuations. Electricity retailers can partially meet the challenge of matching demand and volatile supply by shifting power demand according to the fluctuating supply side. The necessary technology infrastructure such as Advanced Metering Infrastructures for this so-called Demand Response (DR) has advanced. However, little is known about the economic dimension and further effort is strongly needed to realistically quantify the financial impact. To succeed in this goal, we derive an optimization problem that minimizes procurement costs of an electricity retailer in order to control Demand Response usage. The evaluation with historic data shows that cost volatility can be reduced by 7.74%; peak costs drop by 14.35%; and expenditures of retailers can be significantly decreased by 3.52%.
Energy Policy 02/2014; 65:359–368. DOI:10.1016/j.enpol.2013.10.012 · 2.58 Impact Factor
[Show abstract][Hide abstract] ABSTRACT: Auctions are important tools for resource allocation and price negotiations, while combinatorial auctions are perceived to achieve higher efficiency when allocating multiple items. In the recent decade, many auction designs are proposed and proven to be efficient, incentive compatible, and tractable. However most of the results hinge on quasi-linear preference bidders with ultimate patience, which is not quite realistic. In reality human bidders cannot engage in hundreds of auction rounds evaluating thousands of package combinations simultaneously. They either withdraw early or bid only on a limited subset of valuable packages. In this paper, we introduce bid ranges, with an additional sealed-bid phase before a Combinatorial Clock auction for information elicitation. With range information, the auction can start at higher prices with fewer rounds, and bidders are informed with the most relevant packages. Our design reduces the complexity both for the bidders and auctioneer, and is verified with computational simulations.
Proceedings of the 2014 47th Hawaii International Conference on System Sciences; 01/2014
[Show abstract][Hide abstract] ABSTRACT: Speculative bubbles are commonly referred to situations where stock prices considerably deviate from their fundamentals until the bubbles bust. Bursting of bubbles such as the dot-com or U.S. housing bubble is very costly, so there is a need for mechanisms to detect them. In this paper, we attempt to predict when bubbles may bust using the sentiment of news announcements. Accordingly, we first try to understand how news reception evolves depending on the market phase (boom or bust). The probability of bubble bursts are calculated on the basis of a Markov-regime switching model. The approach is applied and validated using the oil market which appears to be one of the most important markets in the globalized world. Our methodology can be similarly extended to other markets such as gold or wheat.
Proceedings of the 2014 47th Hawaii International Conference on System Sciences; 01/2014
[Show abstract][Hide abstract] ABSTRACT: Flash crashes, perceived as sharp drops in market prices that rebound shortly after, have turned the public eye towards the vulnerability of IT-based stock trading. In this paper we explain flash crashes as the result of actions made by rational agents. We argue that information technology, which has long been associated with competitive advantages, may cause ambiguities with respect to the game form that give rise to a Hyper game. We employ Hyper game Theory to demonstrate that a market crash constitutes an equilibrium state if players misperceive the true game. Once the ambiguity is resolved, prices readjust to the appropriate level, creating the characteristic flash crash effect. We also discuss endogenous and exogenous mechanisms that may alleviate the threat of a flash crash and present possible options for future research.
Proceedings of the 2014 47th Hawaii International Conference on System Sciences; 01/2014
[Show abstract][Hide abstract] ABSTRACT: This paper applies novel sentiment analyses to Reuters news to study stock and CDS traders' differential interpretations of financial news. We construct sentiment measures to identify which news content influences investors' behavior and create dynamic word lists that reflect the divergent viewpoints of CDS and equity investors. We find that (1) equity and CDS traders focus on different content within the same news; (2) traders particularly disagree with respect to news concerning debt topics, especially regarding M&A activity; (3) the subprime crisis impacted debt news content and altered the typical inverse relationship between equity and CDS markets on news days.
[Show abstract][Hide abstract] ABSTRACT: In this paper we show how IS research can contribute to two major issues many societies face today – dealing with the decreasing availability of easily accessible fossil fuels and the threat of climate change. We introduce an IS artifact that uses historical data on energy consumption, energy generation, and mobility behavior to unleash synergies between residential photovoltaic panels and electric vehicles. Using real world data we show in simulations that our residential information system allows households to decrease annual net energy costs by up to 68 percent. This results in an earlier break-even point of green technologies and clears a path towards a more sustainable society. Following up with an investment analysis we show that the information system has the same financial impact as the subsidy for plugin-hybrid electric vehicles by the U.S. federal government following the American Recovery and Reinvestment Act – at no cost to the taxpayer.
International Conference on Information Systems 2013, Milan; 12/2013
[Show abstract][Hide abstract] ABSTRACT: Information Systems play an important role in achieving sustainable solutions for the global economy. In particular, Information Systems are inevitable when it comes to the transition from the "current" to the "smart" power grid. This enables an improved balancing of both electricity supply and demand, by shifting load --- based on the projected supply gap and electricity prices --- on the demand side smartly. As this requires a specific Information System, namely a Demand Response system, we address the challenge of designing such a system by utilizing the design science approach: determining general requirements, deducing the corresponding information requirements, analyzing the information flow, designing a suitable Information System, demonstrating its capability, and, finally, evaluating the design. The design process is reiterated fully until a viable solution, i.e. an IS artifact, has been developed. This paper describes both the design process as such and the final IS artifact. Moreover, we summarize our lessons learnt from using and adopting the design science approach within this practical, bottom-up case study.
Proceedings of the 8th international conference on Design Science at the Intersection of Physical and Virtual Design; 06/2013
[Show abstract][Hide abstract] ABSTRACT: This paper investigates whether news momentum can predict medium-term stock index developments. News momentum can be built by aggregating tone of news over the past weeks. We find that news momentum can predict future stock price developments and establish profitable trading strategies that beat buy-and-hold and momentum benchmarks. Trades are issued for significant changes in momentum between current and prior weeks. We ensure stability of our results by using two different news data sets and by analyzing both different investment horizons and aggregation times for our news momentum. Compared to intraday news trading, medium-term momentum trading allows higher investment volumes and can contribute to complex investment decisions also incorporating other qualitative and quantitative factors.
[Show abstract][Hide abstract] ABSTRACT: Decision support systems play an increasingly important role in disaster management research. Coordination of rescue units during disaster response is one of the many areas which may benefit from this development. Time pressure, resource shortages, different capabilities of rescue units and the interdependence of scheduling and allocation tasks belong to the key challenges which emergency operation centers have to cope with. This paper proposes a non-linear optimization model and suggests a Monte Carlo-based heuristic solution procedure. We computationally benchmark our heuristic with a procedure that is applied in practice. Results of our study show that the Monte-Carlo heuristic is superior to the state-of-the art approach in terms of aggregated harm by up to 40%. However, our simulations also reveal that the time our heuristic needs to process medium-sized instances (100 incidents, 50 rescue units) on a PC is a few hours and that more powerful real-time computing capabilities are required.
System Sciences (HICSS), 2013 46th Hawaii International Conference on; 01/2013