Scientific Libraries in Information Markets
Scopes for Organisation and Recommendations for Operation
The project is divided into six subprojects, each of them analysing different aspects of the objective target:
- Subproject A - Market Models and Market Strategies
- Subproject B - Infrastructures
- Subproject C - Recommender Systems
- Subproject D - Payment Systems
- Subproject E - Simulation
- Subproject F - Business Models and Pricing
Subproject A - Market Models and Market Strategies
Within the scope of Subproject A the information market has been examined with emphasis on strategic positionings options for scientific libraries. Therefore the market has been divided into relevant activities and main processes, the positioning of market participants examined and the total market trends resulting from the digitalization analyzed in consideration of the flow of goods, money and information.
As a summary the positive starting point of scientific hybrid libraries for further strategic development could be demonstrated. But also the necessary concentration on core competences has been shown to not stay behind commercial competitors. In order to reach this goal, legal and financial basic conditions have to be set up.
Subproject B - Infrastructures
The task of Subproject B was to develop a platform for electronic trading integrating not only libraries but also other providers of scientific literature. The platform is supposed to enable the employment and the exploration of new services and was orientated to the model of an open and dynamic market. Developing services to be integrated into the system, emphasis was put on the linkage and location of providers, the result integration and the support of customers.
The outcome of this work is the cross programming language and cross platform multiagent sytem UniCats (Universal Integration of Catalogues based on an Agent supported Trading and Wrapping System). Purpose of the provider linkage, realized with provider agents, is the finding of suitable providers for a customer's inquiry. The selection can be based not only on subject or monetary criteria but also on the customer's location. The integration of results, realized with integration agents, gives the opportunity to send inquiries to several providers and to combine the results into a uniform list. Finally customer agents represent customers in the system and support them with the planning of inquiries.
Subproject C - Recommender Systems
Within the scope of Subproject C recommender services as an example for innovative services for scientific libraries have been studied. The acceptance of these systems can be demonstrated by their successful employment by commercial providers such as the internet book trader Amazon. An analysis has revealed that great advantages can be expected deploying recommendation services in scientific libraries with their enormous stock of high quality documents.
In this project a stochastic user behaviour model based on Ehrenberg's Repeat Buying theory has been developed. By means of this model recommendations based on the behaviour of library users can be determined. A recommender software has been developed and a recommender service has been integrated into the OPAC (Online Public Access Catalogue) of the University Library of Karlsruhe. It is now after the end of the project open to all library users as a standard service. It covers the compound catalog of the library system of Southwest Germany with 15 million documents in 23 libraries and is the worlds largest recommender system of its kind. With the help of this service literature recommendation can be requested, which are based on the behaviour data of users. These recommendations are other documents within the system (possibly from other libraries), which are more frequently used together than usual as can be proven by statistical tests. As the data basis is constantly expanded through usage of the OPAC, the number of available recommendations increases continously. The catalog is used by approximately one million users a month.
Several evaluations by user statistics and user surveys were carried out, resulting in a very positive response. In addition the suitability of the system and the generated statistics for the support of the stock management could be shown.
Subproject D - Payment Systems
An other important aspect of information markets for scientific literature supply is the availability of payment systems, as finacial transactions in business processes have to be taken into consideration. Therefore the platform UniCats developed in Subproject B has been expanded to enable the performance and the record of financial transactions.
In this context invoice agents, which are informed of all financial transactions by other agents, are fundamentally important. They build an independent authority for the proof of arranged transactions by keeping accounts for registered agents. Payment flows can be booked in these accounts so that any payments within the UniCats system can be realised without the exchange of money. To enable financial transactions with customers and suppliers, payment agents support a connection to external finance mediators for traditional and electronic payment. Therefore a media break can be avoided as payments of customers to suppliers can be processed within the UniCats system.
In order to carry out finacial transactions, security mechanisms had to be implemented into the UniCats system. The encrypted exchange of data can be done by the implementation of SSL (Secure Socket Layer) on the level of the multi agent system.Therefore the security of the UniCats system corresponds to this common protocol used for electronic payments. Finally identity cards administrated by special agents were introduced to check the identity and authorisation of agents and customers.
Subproject E - Simulation
The aim of Subproject E was the simulation of market strategies and infrastructures that were worked out in other subprojects. Therefore an open market as it was realised in the UniCats system has been modeled and it has been tested, if the model of the electronic market meets the requirements. The simulation experiments could confirm that the model of the electronic market is suited for a great number of suppliers and customers and more then 1000 agents as well as it is able to adapt the equilibrium in case of disturbance (for example due to breakdown of agents).
Beside the common market model further simulation experiments deal with the scientific literature supply for students. It is based on an abstract mathematical model with explicit input and output parameters, which describes dependences between the individual market participants. By adapting the input parameters, any initial situation could be simulated in this environment
Subproject F - Business Models and Pricing
Within the scope of Subproject F the acceptance of priced services and price elasticities within these services were examined. Therefore in this project the new document delivery service LEA for students (Local Electronic Document Delivery System) was introduced at the University Library of Karlsruhe. This service delivers electronic documents, which have been ordered at a web interface, to students. Service data for the analysis of price elasticities were gathered by prices changes dependend on demand and service capacities as well as special discounts.
The most important observations can be summed up as followed. Prices, which are accepted by students, are far below todays cost-covering bound. This is mainly a consequence of the actual copyright law according to which it is not legal to store digital information for further inquiries. Therefore a repeated inquiry of the same article leads to the very cost-intensive repeated manual scan of the same paper document. Although the service has been strongly advertised, the number of users increases very slowly and the demand is limited mainly to students, who are working on their Diplom (final thesis). The diffusion time of the sevice among students lies beyond the end time of the project. However, distinctive price elasticities could be proved as a low price has only slight effects on the demand whereas the demand drops to zero at a significantly higher price because the students still have the opportunity to manually copy the paper documents on their own.