Multi-agent Organization for Hiberarchy Dynamic Evolution

With increasingly dynamic operating environment and user requirements, software adopts a unified strategy to achieve the different levels of evolution, a fact which reduces the flexibility and efficiency. So, in this paper, a method with agent technology is proposed to support the hiberarchy evolution of both the function and service levels. Precisely, a multi-agent organization is proposed to separate the calculation and collaboration logics of software which are corresponding to the different levels of evolution. To achieve the function-level evolution, an adaptive agent model with knowledge reasoning provides the software an ability to dynamically modify the calculation logics. With the adjustment of the collaboration logics, the multi-agent organization can make it convenient for the software to deal with the service-level evolution. Finally, a case study of air defense simulation system and some test metrics indicates that the proposed multi-agent organization can effectively support the hierarchy evolution.