Decision support systems configuration based on knowledge-driven automated service composition: requirements and conceptual model

Abstract

The development of decision support systems is a complex, multi-stage process that requires considerable effort both methodically and technically. At the same time, with the proliferation of service-oriented architecture (and, more recently, the microservice approach based on it), technical development of such systems is often largely reduced to the construction of compositions of existing services implementing information processing functions, so that eventually the corresponding composition provides the necessary functionality of the decision support system. The paper proposes an approach to the construction of configurable service-oriented decision support systems based on the automated service composition, which will significantly reduce the effort required to develop such systems. Particularly, following results are presented: a) functional framework of different types of decision support systems, typical structures and patterns applicable in different types of decision support systems; b) a set of requirements for configurable service-oriented decision support systems and their main components (tools for describing services, and the purposes of the composite service, methods and algorithms for implementing the composition); c) a conceptual model of a configurable service-oriented decision support system.