Complex open-system design by quasi-agents: process-oriented modeling in agent-based systems

  • Authors:

  • Behzad Bastani;David Greaves

  • Affiliations:

  • University of Cambridge, Cambridge, UK;University of Cambridge, Cambridge, UK

  • Venue:
  • ACM SIGSOFT Software Engineering Notes
  • Year:
  • 2009

Quantified Score

Hi-index 0.00

Visualization

Abstract

Drastically increasing involvement of Computer Science in different aspects of human life and sciences and the reciprocal dependency these sciences have developed on Computer Science and technology have deployed extremely challenging grounds for software architecture and design as a discipline. The requirements related to the management of complexity which is in the essence of these new domains combined with sizes which are orders of magnitude larger than the conventional business applications necessitate development of new paradigms. Since we are in fact beyond the age of writing one program by one group which takes care of one type of issue for one class of users, the new paradigms should guarantee some type of technical pluralism which allows indefinite number of people addressing indefinite aspects of complex clusters of issues in an ongoing effort over indefinite amount of time. The technical basis should provide for easy and ideally automated integration of all such efforts. This means the capability of random program design and integration based on a supporting and unifying conceptual framework. In contributing to these principles, Nuclear-Process Oriented Analysis and Modeling (NPOAM) [3] presented the capability of random modeling and design while establishing itself on the supporting framework of Abstraction-oriented Frames [4]. In an ongoing research to further modularize and streamline this methodology, this paper presents the idea and method of application of NPOAM to agent-base systems. The accomplishment of this goal is substantiated through a double implementation effort, one in an agent simulation environment and the other through the use of industrial strength modeling and application development tools. This paper also extends the agent-oriented framework to propose a new concept named "quasi-agents" which is essentially related to mostly deterministic environments and offers examples of quasi-agents in implementation.