| Multi-agent systems are already a focus of studies for more than 25 years. Despite substantial effort of an active research community, modeling of multi-agent systems still lacks complete and proper definition, general acceptance, and practical application. Due to the vast potential of these systems e.g. to improve the practice in software and to extent the applications that can feasibly be tackled, this book tries to provide a comprehensive modeling language - the Agent-Modeling Language (AML) - as an extension of UML 2.0, concentrating on multi-agent systems and applications.
Today, in 2007, the field of study known as multi-agent systems has been in existence for more than 25 years. However, only during the mid-1990s did the field begin to draw widespread attention as the hype-curve approached its zenith. Now as the first decade of the 21st century begins to wane, we find the field ever more active with branches in a myriad of disciplines and aspects of multi-agent systems theory and engineering in use throughout multiple application domains and business sectors. However, one important aspect of multi-agent systems that still lacks complete and proper definition, general acceptance and practical application, is that of modeling, despite the substantial efforts of an active research community.
In short, software agents are a domain-agnostic means for building distributed applications that can be used to create artificial social systems. Their facility for autonomous action is the primary differentiating property from traditional object-oriented systems and it is this aspect that most strongly implies that standard UML is insufficient for modeling multi-agent systems.
The focus of this book is thus on an approach to resolving this insufficiency by providing a comprehensive modeling language designed as an extension to UML 2.0, focused specifically on the modeling of multi-agent systems and applications. This language is AML—the Agent Modeling Language—the design of which is informed by previous work in this area while explicitly addressing known limitations relating to managing complexity and improving coverage and comprehension. |