Using Relative Complexity Measurement Which from Complex Network Method to Allocate Resources in Complex Software System's Gray-Box Testing

Abstract

Software testing costs would be reduced if we could find which classes or models of software are more complex and thus more likely to have defects. This paper defines a method which uses the software complexity measurement from complex network method to allocate resources in complex software system for gray-box testing. It presents the way of mapping the software codes to complex network and some metric's effects in software network, such as degree and degree of ripple. By using these metrics, it defines the complexity measurement to complex software network. The work about C++ complex software system shows how to use this method.