An agent-based framework for software evolution
Abstract
In this work, we extend the current field of object-oriented software engineering by deriving a more complete set of object-oriented meta-patterns. We use our meta-patterns to decompose object-oriented design patterns and separate the concerns that cause shortcomings in the area of software evolution. We then extend our object-oriented meta-patterns by using the agent concepts to reduce the cost and risk associated with each cycle of software evolution. We finally compose a set of agent-based patterns from the agent-based meta-patterns and demonstrate that they are a worthwhile improvement to the field of software evolution.
Keywords
Thesis