Research And Implementation Of A Mechanism For Agent-based Software Self-adaptive Integration Evolution

As the scale of software keeps growing and the dynamics of operating environmentbecome increasingly prominent, research on software adaptive and dynamic evolutionin an integrated way has important theoretical and engineering significance in reusingexisting resources, constructing high availability software systems, prolonging softwarelife cycle, improving the adaptive capability of software to changing operatingenvironment and reducing software development and maintenance costs. Agent possessthe characteristics of autonomy, intelligence and adaptability that are desirable fordescribing systems which has autonomic members with flexible interaction requirementin changing environment. This paper applies Agent technology to software integrationevolution, and proposes an Agent-based software adaptively integration evolutionmechanism to provide support for software to evolve adaptively in an integrated wayaccording to environment changes.First of all, the paper puts forward an Agent-based software adaptive integrationevolution framework and a supported Agent model based on the correspondingcharacteristic and requirement analysis. Then it focuses on the environment modelwhich drives software to evolve adaptively, including the environment classification anddefinition, and designs environmental change perception mechanism based on eventtriggering. And by introducing ECA to describe evolution rule, the paper designs thestrategies for software adaptively evolving according to inside and outsideenvironmental changes. Based on the designs, the paper presents the realization of themechanism. Finally, by taking a simulation system with typical integration evolutioncharacteristics as a test case, the paper carries out the software adaptive integrationevolution experiment, and the result indicates validity and efficiency.