O-MaSE: An Extensible Methodology for Multi-agent Systems
As the complexity of software systems continues to grow, the multi-agent systems approach has been proposed as an approach to handling this complexity. A key factor to the use of multi-agent systems in real systems is the existence of industrial strength
- PDF / 736,041 Bytes
- 19 Pages / 439.36 x 666.15 pts Page_size
- 82 Downloads / 183 Views
O-MaSE: An Extensible Methodology for Multi-agent Systems Scott A. DeLoach
Abstract As the complexity of software systems continues to grow, the multi-agent systems approach has been proposed as an approach to handling this complexity. A key factor to the use of multi-agent systems in real systems is the existence of industrial strength development methodologies. The Organization-based Multiagent Software Engineering (O-MaSE) methodology was created in response to this realization. O-MaSE integrates a set of concrete technologies aimed at making multi-agent technology available to industry and facilitating widespread acceptance. Specifically, O-MaSE was created as a customizable methodology that can be adapted and extended for a wide variety of uses. Keywords Agent-oriented methodology • Method engineering • Metamodel • Software analysis • Software design
1 Introduction One of the discoveries related to software methodologies over the last several decades is that there is not a single methodology that will work for all software development projects. This is true due to the vast number of different software applications as well as the increasingly large number of paradigms used to develop software. To develop complex, distributed, and adaptive software solutions, multiagent systems (MAS) have been proposed as a promising approach to meet these difficult demands [1]. However, even within a given paradigm such as MAS, a single methodology or approach will never be sufficient given the broad range of applications for which MAS are applicable. Thus, either several unique methodologies must be created to handle a broad range of MAS applications or an approach must
S.A. DeLoach () Kansas State University, Manhattan, KS, USA e-mail: [email protected] O. Shehory and A. Sturm (eds.), Agent-Oriented Software Engineering, DOI 10.1007/978-3-642-54432-3__9, © Springer-Verlag Berlin Heidelberg 2014
173
174
S.A. DeLoach
be found to tailor methodologies to their proposed applications. One approach to supporting the ability to tailor methodologies to a specific application is situational method engineering (SME) [2]. Henderson-Sellers [3] was one of the first to argue that SME is the key to creating flexible, industrial strength methodologies as it allows the development of standard approaches that are widely used and accepted while continuing to allow the methods and processes of those approaches to be modified and extended. SME provides this flexibility by allowing method engineers to construct methods and processes1 (a.k.a. methodologies) from a set of existing method fragments. There has been some work at unifying the disparate MAS methodologies. For example, the Agent-Oriented Software Engineering Technical Forum Group (AOSE-TFG) was created to help find a way to allow MAS methodologies to move toward interoperability by creating a common metamodel of MAS concepts [4, 5]. More recently, the IEEE FIPA Design Process Documentation and Fragmentation working group has been working to standardize the definition of existi
Data Loading...
