Title: A knowledge-driven approach for specifying the requirements of multi-agent system

Authors: Vibha Gaur; Anuja Soni

Addresses: Department of Computer Science, New Academic Block, University of Delhi, Delhi-1100017, India ' Department of Computer Science, New Academic Block, University of Delhi, Delhi-1100017, India

Abstract: Agents exhibit a high degree of cooperation to achieve designated goals that may not be feasible otherwise. In such a cooperative environment, the specification of the static and dynamic requirements of agents is a prime concern. Knowledge is the acquaintance with the facts or principles for extracting the non-trivial and potentially useful information from data. This work comprehends the knowledge in terms of decision table, indiscernible relation, information system, equivalence classes and reduction of rules. To specify the static requirements, this work proposes a coding scheme that assists the developers in obtaining the equivalence classes and enhances the readability and maintainability of the requirements. To specify the dynamic requirements, this work utilises the knowledge rule-base and visual state flow (SF) charts. Rule-base consists of the rules to identify appropriate service provider agents whereas visual state flow (SF) charts are used to comprehend the inter-agent dependency requirements and also enhance the chances of detecting inconsistent requirements.

Keywords: software requirements specification; SRS; multi-agent systems; MAS; coding scheme; knowledge rule base; visual state flow charts; static requirements; dynamic requirements; agent-based systems; cooperation; inter-agent dependency.

DOI: 10.1504/IJBIS.2015.069722

International Journal of Business Information Systems, 2015 Vol.19 No.3, pp.300 - 323

Published online: 30 May 2015 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article