Software Engineering for Multi-Agent Systems II

Advances in networking technology have revitalized the investigation of agent technology as a promising paradigm for engineering complex distributed software systems. Agent technology has been applied to a wide range of application domains, including e-commerce, human-computer interfaces, telecommunications, and software assistants.

Multi-agent systems (MASs) and their underlying theories provide more natural support for ensuring important properties such as autonomy, mobility, environment heterogeneity, organization, openness, and intelligence. As a consequence, agent-based systems are likely to provide new approaches to dealing with the complexity of developing and maintaining modern software.

However, developing robust, large-scale agent-based systems will require new software engineering approaches. There are currently many methods and techniques for working with individual agents or with systems built using only a few agents. Unfortunately, agent-based software engineering is still in its infancy, and existing software engineering approaches are unable to cope with large MASs.

The complexity associated with a large MAS is considerable. When a huge number of agents interact across heterogeneous environments, various phenomena occur that are not as easy to capture as when only a few agents are working together. As multiple software agents are highly collaborative and operate in networked environments, they have to be context-aware and deal with environmental uncertainty. This makes their coordination and management more difficult and increases the likelihood of exceptional situations, such as security holes, privacy violations, and unexpected global effects.

Moreover, as users and software engineers delegate more autonomy to their MASs and place more trust in their results, new concerns arise in real-life applications.