Structured development of problem solving methods

Abstract

. Problem solving methods (PSMs) are domain-independent reasoning components, which specify patterns of behavior which can be reused across applications. While the availability of extensive PSM libraries and the emerging consensus on PSM specification languages indicate the maturity of the field, a number of important research issues are still open. In particular, very little progress has been achieved on foundational and methodological issues. Existing libraries of PSMs lack a clear theoretical basis and only provide weak support for the method development process, usually in the form of informal guidelines. In this paper we will address these issues by illustrating a framework which characterizes PSMs in terms of problem commitments , problem-solving paradigms and domain assumptions . This framework provides i) a theoretical foundation for situating PSM research and individual PSMs, as well as ii) an organization which allows us to characterize method development and selection as a process of navigating through a three-dimensional space (defined by the three components of our framework). Individual moves through this space are specified by means of adapters . In the paper we will illustrate these ideas in detail, with examples taken from parametric design problem solving.

Choreography in IRS-III – Coping with Heterogeneous Interaction Patterns in Web Services

“…IRS-I [4] supported the creation of knowledge intensive systems structured according to the UPML framework [9] and IRS-II [15] integrated the UPML framework with web service technologies. Within IRS-III we have now incorporated and extended the WSMO ontology [17].…”

Collaborative Semantic Web Browsing with Magpie

“…The architecture uses an infrastructure that we have developed, called IRS-II [20], which supports the publishing and invocation of semantic services. A key feature of IRS-II is that it is based on the UPML framework [11] and therefore differentiates between tasks, problem solving methods (generic reasoners) and domain models. The split between tasks and problem solving methods enables us to clearly separate the activities of implementing semantic services from making them available in a fashion where they can be easily invoked with little or no overhead to the end-user.…”

Configuring Online Problem-Solving Resources with the Internet Reasoning Service

“…Our approach sterns from research on knowledge-based system (KBS) development by reuse [1,3,4,6,7,11]. This area has identified domainindependent problem-solving strategies, or problem-solving methods (PSMs) [1,6,7,11,12], that provide standard ways of addressing stereotypical knowledge-intensive problems, or generic tasks [4], such as diagnosis, design, and classification.…”

“…This area has identified domainindependent problem-solving strategies, or problem-solving methods (PSMs) [1,6,7,11,12], that provide standard ways of addressing stereotypical knowledge-intensive problems, or generic tasks [4], such as diagnosis, design, and classification. To foster the reuse of PSMs, structured libraries have been developed, in which the methods are indexed and retrieved for different domains and purposes [1, 3,6,[11][12][13].…”

“…To foster the reuse of PSMs, structured libraries have been developed, in which the methods are indexed and retrieved for different domains and purposes [1, 3,6,[11][12][13]. Furthermore, comprehensive frameworks take a PSM-centered view to support KBS development from reusable components [7,11,14,17]. This process involves modeling a user problem as a generic task, and configuring an appropriate PSM, by integrating the generic components with domain-specific information.…”