Research On Neural Component Model And Software Architecture Dynamic Evolution

Software evolution refers to the process of evolution of achieving the desired form by the software changes. It includes static and dynamic evolutions. Because of the advantages of an ongoing availability the dynamic evolution of software engineering research has become a hot spot. But the dynamic evolution is more complex and more difficult than static evolution to handle in technology. The complexity of modern software determines the research of dynamic evolution should start with the macro level. To support SA (software architecture)-based dynamic evolution, this article makes some systematic research on component, software architecture framework and dynamic configuration etc.Based on neuron conduction theory, we make a neural component information analysis model (NCIAM) by imitating the electrical signal conduction of the neuron, and make a neural component model with it. In neural component, information conduction between one particular component and other components of is modeled as an abstract type of the simulations of nerve conduction. The component's other traditional features are modeled into a new component body; the point of interaction between the components are modeled as information transduction linkers.On the basis of the neural component model, we continued the research on the software architecture of dynamic evolution. Using the Multi-Agent system design idea we made a neural component-based dynamic software architecture model NDSA (Neuron component model-based DSA). We defined the basic elements of the model and explained the composition of the model structure. And we analyzed the local and global evolution of the model. On the basis of the model, from the view of large-grained, we described and analyzed the software architecture framework of the dynamic evolution based on the model NDSA.Ultimately, in order to make sure the dynamic evolution based on neural component software architecture is more perfect more efficient implementation of dynamic evolution, we introduced the theory of a reflex configuration model. According to the model we established a dynamic configuration system based on the model.