| The complexity of software, on the one hand, is reflected by the scale and variation of internal control. On the other hand, with popularization of WWW and development of Internet technology, the application environment of user-centric, service-oriented networked software takes new complexity to software system. It is because of both internal and external factors make software indigestible, nondeterministic, uncontrollable and lead to inconvenience in the aspects of design, development, maintenance and management.The maturation of complex network theory and followed new discoveries provide tool to understand the artificial self-adaptive complex system. By taking the software as complex network, analyzing the structural feature, dynamic behavior and evolution principle as a whole, revealing some hidden patterns and universal laws, are useful to understand internal essential characteristics comprehensively, and have very important significance in evaluating software complex and optimizing system structure, improving availability and stability of software systems, etc.This thesis researches the dynamic behavior of the software execution network based on complex network theory, it includes four main parts as follows:(1) Defining software execution network based on function call relationship, and giving the construction methods of software execution network(2) Researching the execution network of three open-source software systems. Mainly including:measuring the structure of execution network, and comparing with the static function network; analyzing the behavior role of key nodes in execution network; analyzing the structure evolution of execution network.(3) Proposing weighted execution network model and analyzing its structure features.(4) Establishing user behavior model based on Markov random process, mining the frequent operations of large number of users and generating key execution network, measuring and analyzing this network. |
PDF Full Text Request