Suitability assessment framework of agent-based software architectures

Abstract

Context: A common distributed intelligent system architecture is Multi Agent Systems (MASs). Creating systems with this architecture has been recently supported by Agent Oriented Software Engineering (AOSE) methodologies. But two questions remain: how do we determine the suitability of a MAS implementation for a particular problem? And can this be determined without AOSE expertise? Objective: Given the relatively small number of software engineers that are AOSE experts, many problems that could be better solved with a MAS system are solved using more commonly known but not necessarily as suitable development approaches (e.g. object-oriented). The paper aims to empower software engineers, who are not necessarily AOSE experts, in deciding whether or not they should advocate the use of an MAS technology for a given project. Method: The paper will construct a systematic framework to identify key criteria in a problem requirement definition to assess the suitability of a MAS solution. The criteria are first identified using an iterative process. The features are initially identified from MAS implementations, and then validated against related work. This is followed by a statistical analysis of 25 problems that characterise agent-oriented solutions previously developed to group features into key criteria. Results: Key criteria were sufficiently prominent using factor analysis to construct a framework which provides a process that identifies within the requirements the criteria discovered. This framework is then evaluated for assessing suitability of a MAS architecture, by non-AOSE experts, on two real world problems: an electricity market simulation and a financial accounting system. Conclusion: Substituting a software engineer's personal inclination to (or not to) use a MAS, our framework provides an objective mechanism. It can supplant current practices where the decision to use a MAS architecture for a given problem remains an informal process. It was successfully illustrated on two real world problems to assess the suitability of a MAS implementation. This paper will potentially facilitate the take up of MAS technology. © 2012 Elsevier B.V. All rights reserved. This journal article is available at Research Online: http://ro.uow.edu.au/eispapers/600 AbstractContext: A common distributed intelligent system architecture is Multi Agent Systems (MAS). Creating systems with this architecture has been recently supported by Agent Oriented Software Engineering (AOSE) methodologies. But two questions remain: how do we determine the suitability of a MAS implementation for a particular problem? And can this be determined without AOSE expertise?Objective: Given the relatively small number of software engineers that are AOSE experts, many problems that could be better solved with a MAS system are solved using more commonly known but not necessarily as suitable development approaches (e.g. object-oriented). The paper aims to empower software engineers, who are not necessarily AOSE experts, in deciding whether or not they...

Understanding Cyber Security In Health Sector

“…The authors of [66] suggest an ECC-based encryption scheme that makes use of homomorphism encryption. The GASONeC method, which used for the encryption pattern, which had been built on it [37] [41]. ECC was formerly applied to swap open and private keys as a result of its capability to provide a high level of security while having a key size of just 176 bits [8] [9].…”

“…ECC was formerly applied to swap open and private keys as a result of its capability to provide a high level of security while having a key size of just 176 bits [8] [9]. The authors of the article [67] provided the design and implementation of a lightweight hub validation protocol, which covers hub enlistment, hub confirmation, and important foundation stages [37] [45]. However, neither a digital signature technique nor any open key cryptography is used in any way [37].…”

Efficient Software Architecture Pattern for Accelerator Based Computing

“… Distributed Agent based architecture  Socialistic architecture  Master-slave architecture Distributed agent based software architectures have been presented in [6] and was successfully exploited for various distributed and parallel computing applications [7][8] and [9]. Further studies were carried out on this architecture in [10] and [11].…”

A Refined Metric Suite for a Multi Agent System

“…For creating Multi Agent System (MAS) Agent Oriented Software Engineering (AOSE) is supported [11]. But there arises two questions that is the first is how do we determine the suitability of MAS implementation of a particular problem and the next is how this can be done without an AOSE expert?…”

“…The software agent methodologies are PROMETHEUS, GAIA, OperA, O-MaSE, Tropos and GORMAS [11] [7]. GAIA is a methodology for agent-oriented analysis and design.…”