Business Systems and Knowledge Modelling Lab

Research Project

Multi-Layered Agent-based Architecture for Modeling Intelligent Accounting Systems

Research Goal

1. Design and develop multi-layered agent based architecture for intelligent accounting systems. The architecture facilitates definition of the soft computing artifacts in the context of tasks of the practitioner/accountant.
2. Measure and optimize the performance and quality of the solution of soft computing technologies employed by intelligent accounting systems.

Acknowledgements
This project is being done in collaboration with procurement and finance divisions (users of SAP) of Latrobe University

Problem Space
The wide and increasing needs of adaptable, flexible and predictive solutions for many accounting and auditing problems point out the importance of using design methodologies and implementation technologies with high ability to learning, adaptation and evolution. Soft computing is one of the most relevant answers to such needs: neural networks, fuzzy logic, and genetic/evolutionary algorithms are being used to model a range of problems in accounting and auditing including bankruptcy prediction, learning auditor’s decision making behaviour in fraud settings and forensic accounting.

On the other hand, accuracy and uncertainty issues must be carefully considered in these applications since the quality of the solution greatly relies on them. Up to now, analysis and experiments have been performed by researchers and practitioners mainly to understand the underlying technologies and their applications in accounting and auditing, but without any specific focus on the mandatory need to measure and optimize the quality of solution provided by these technologies. These aspects are vital to ensure the correct and effective use of intelligent accounting systems employing soft computing technologies in industrial environments. Secondly, it is useful to apply these technologies in the context of decision support models of the accountants and auditors.

Team Members
Rajiv Khosla

Publications
[1] R. Khosla “Human-Centred Intelligent Systems,” in the International Journal of Fuzzy Systems and Intelligent Informatics, Vol. 15, No. 4, pp. 379-390, 2003

[2] R.Khosla, E Damiani and W. Grosky Human-Centered e-Business, Kluwer
Academic Publishers, Massachusetts, USA, March 2003, 315 pages