MIAC: Methodology for Intelligent Agents Componentware


The DAI-Labor at Technische Universität Berlin has concerned itself with Agent-Orientation for more than 15 years. Numerous projects have studied the application of the Agent-Oriented Programming paradigm to real-world problems in diverse application areas such as automotive, entertainment, production control, connected living, office and software engineering (as applying research results on the field itself). Distributed artificial intelligence and agent technology promise to be the key factors to engineering large-scale, distributed and complex software systems and services. They make it easy to understand large problem fields, and provide means to their solution, based on natural, plain and intelligible metaphors. This thesis integrates three keystones of Agent-Oriented Software Engineering: an agent-oriented framework, the agent-oriented methodology, and an agent-oriented and agent-based tool suite. The result is submitted to an iterative and incremental process itself in order to streamline these building blocks for fast and efficient development of agent-based applications and services. The outcome of this thesis are products that have been developed, applied and evaluated in projects together with students of the TU Berlin, other Competence Centers of the DAI-Labor and also with partners from different industrial sectors. While many agent-oriented methodologies and tools usually reflect theories and laboratory experiments, the framework, methodology and tools presented in this thesis have developed market-ready prototypes and competitive solutions that have also successfully shown their value in different programming contests.

Dissertation, Technische Universität Berlin, Berlin, Germany