Consistency-based diagnosis of configuration knowledge bases

Abstract

Configuration problems are a thriving application area for declarative knowledge representation that currently experiences a constant increase in size and complexity of knowledge bases. Automated support of the debugging process of such knowledge bases is a necessary prerequisite for effective development of configurators. We show that this task can be achieved by consistency-based diagnosis techniques. Based on the formal definition of consistency-based configuration we develop a framework suitable for diagnosing configuration knowledge bases. During the test phase of configurators, valid and invalid examples are used to test the correctness of the system. In case such examples lead to unintended results, debugging of the knowledge base is initiated. Starting from a clear definition of diagnosis in the configuration domain we develop an algorithm based on conflicts. Our framework is general enough for its adaptation to diagnosing customer requirements to identify unachievable conditions during configuration sessions. A prototype implementation using commercial constraint-based configurator libraries shows the feasibility of diagnosis within the tight time bounds of interactive debugging sessions. Finally, we discuss the usefulness of the outcomes of the diagnostic process in different scenarios.

Empirical Knowledge Engineering: Cognitive Aspects in the Development of Constraint-Based Recommenders

“…V C = {vc 1 :user.knowledge-level, vc 2 :user.willingness-to-take-risks, vc 3 :user.durationof-investment, vc 4 :user.advisory-wanted, vc 5 :user.funds-available, vc 6 :user.directproduct-search, vc 7 4 :product.min-invest-period, vprod 5 4 : user.willingness-to-take-risks = low → product.risklevel = low, cf 5 : user.willingness-to-take-risks = medium → product.risklevel = low ∨ product.risklevel = medium, cf 6 : user.willingness-to-take-risks = high → product.risklevel = low ∨ product.risklevel = medium ∨ product.risklevel = high, cf 7 : user.knowledge-level = beginner → product.risklevel ≠ high, cf 8 : user.type-low-risk-products = savings → product.name = savings, cf 9 : user.type-low-risk-products = bonds → product.name = bonds}…”

“…These environments provide intelligent modeling functionalities, for example, the automated fault detection in recommender knowledge bases. The principle of such a debugging functionality is that domain experts provide examples for the intended input/output behavior of the recommender knowledge base and a diagnosis component automatically determines the set of potentially faulty recommendation constraints (for details see [4]). …”

Web-Based Commerce in Complex Products and Services with Multiple Suppliers

“…Therefore, we employ an ontological layer, based on a logic theory of configuration described in [6] that enables communication by mapping the specific representations onto more general ontological concepts from the configuration domain (compare [17]). Problem-solving itself can be compared to the work of [19], where the authors of that publication describe how several agents can build major subassemblies on their own, but these subassemblies must "hook together" in a compatible way.…”

Research Issues in Knowledge-Based Configuration

“…A detailed discussion on the application of model-based diagnosis concepts to the automated debugging of configuration knowledge bases can be found in (Felfernig et al, 2004): A CKB Diagnosis Problem is a triple (DD, E + , E -), where DD is a configuration knowledge base, E + is a set of positive examples (test cases), and E -is a set of negative examples (test cases). The examples are given as sets of logical sentences.…”

“…In fact, significant advances have been made since these early, "rule-based" years: Powerful knowledge representation schemes for configuration problems have been developed (Mittal, S. & Frayman, F., 1989;, a formalization of the problem has been done (Felfernig et al, 2004), the invention of new algorithms was driven by the challenges of the domain (Mittal & Falkenhainer, 1990;Fleischanderl et al, 1988), industrial-strength software libraries are now available (Mailharro, 1998) and knowledge based configurators are nowadays already incorporated in standard business software, see for instance (Haag, 1998).…”