25. July 2016

We congratulate Marie Christin Platenius to her doctorate

Marie Christin Platenius has successfully passed her doctoral examination. She received her doctorate for the topic "Fuzzy Matching of Comprehensive Service Specifications“ under Prof. Dr. Wilhelm Schäfer.

In the last decades, software development turned from monolithic software products towards flexibly combinable software services. Correspondingly, a growing number of service providers started offering their services in world-wide service markets. As a consequence, customers, so-called service requesters, can buy software services from such markets. In general, service requesters buy services that match their functional and non-functional requirements best. Thus, among all services provided in the market, the services that match these requirements best need to be discovered by considering these requirements. For the purpose of this so-called service matching, the requester’s requirements specification and the providers’ service specifications are compared. As a result, the extent to which the provided service specification satisfies the requirements specification is returned.

The described scenario leads to two main challenges. (1) Existing matching approaches deliver inaccurate results because each of them only considers a specific kind of requirements. The combination of multiple matching approaches, however, is a tedious and error-prone task and this task had to be done manually up to now. Challenges include design decisions about control flow and data flow as well as the aggregation of matching results. (2) The more complex the specifications we are dealing with are, the more the probability for (partially) imperfect specifications increases. For example, requesters often have vague requirements or deliver imprecise requirements specifications. Providers often provide incomplete specifications for several reasons including the high specification effort or a lack of technical knowledge. Furthermore, matching approaches themselves can be imprecise on purpose to keep complex matching problems efficiently solvable. In these cases, traditional matching approaches deliver adulterated and deceptive matching results that do not notify the users about the induced uncertainty.

In order to deal with complex service specifications, in this thesis, we propose the idea of matching processes that combine multiple existing matching approaches and aggregate their matching results. Thereby, a variety of different requirements can be covered. For this purpose, a model-driven development framework for comprehensive and configurable service matching, called MatchBox, is introduced. MatchBox simplifies, validates, and partially automatizes complex integration tasks.

In order to cope with uncertainty induced into the matching procedure, we propose concepts for Fuzzy Matching. On the basis of well-defined fuzziness sources and types, the amount of induced fuzziness is quantified and returned as part of an informative matching result that reflects the induced uncertainty to the user. Thereby, fuzzy matching provides valuable information about the quality of the matching result, which improves the decision making of both service requesters and service providers.

We realized our concepts in a prototype that was also used for the evaluation including four case studies. Both researchers and practitioners benefit from the contributions presented by this thesis. By combining multiple research areas in a novel way, this thesis describes and evaluates concepts that go significantly beyond the state of the art in service matching. Thereby, it constitutes an important step to bring service matching into practice.

The doctorate will be published soon at the HNI-press.


back to overview