| The ever-increasing complexity of modern software systems has become a major obstacle to their development and evolution. In response to this, agent-orientation has emerged as a promising paradigm for engineering organizational software. This thesis explores the intersecting territories of agent-oriented software engineering, multiagent systems, agent-based economics, and biological evolution to offer a comprehensive framework guiding agent-oriented software development and evolution. In particular, grounded in the Tropos ontology, the study formally examines the defining properties of "agent" as a software engineering paradigm, and presents the taxonomy of goal relations. Based on these, the Tropos Heuristics for Engineering Multiagent sofwarE ("THEME") is proposed to guide agent-oriented software development, and the Tropos Evolution Modeling Process for Organizations ("TEMPO") is introduced to help agent-based software systems evolve in concert with changes in the socioeconomic environments. THEME and TEMPO are both illustrated and preliminarily evaluated with real-life based case studies. |
