| The research is based on the national science foundation, "The measurement method and system of aerospace embedded trusted software". In order to check the errors in the software system which may be reduce the reliability, this thesis proposes a path-wise automated generation test data system.This thesis focuses mainly on the deriving and solving of system of constraints, then points on the way in how to analyse and reduce the statement and the primary obstacle in solving system of nonlinear constraint(s) among them. The progress of the studies on problem is introdueed in this thesis. The existing approaches to address this problem are classified into four categories:random, static, dynamic and heuristic. The representative methods in each category are introdueed and analyzed. The direction of researeh is also explored. Neelam Gupta et al proposed a method, which is referred to as the Iterative Relaxation Method in this thesis, to address the above problem. By means of analyzing the static and dynamic data dependencies between the statements on W, and constructing predicate slices and input dependency sets, this method linearizes the predicate funetions with linear arithmetic representations. Therefore it can only be used to generate test data for white-box testing. Neelam Gupta et al developed a constraint solving technique by using the Least Square Error Solution. As the process of sloving the problem of test data is based on the choosing of paths, this thesis proposes a decision tree algorithm called ID3 which is used to choose the inputs to get the decision tree of the given program, which we make use of computing the x. And when P executes on x, path W will be traversed.The Iterative Relaxation Method is improved in this thesis by omitting the constructions of predicate slices and input dependeney sets. Furthermore, when the divided differences are used to approximate to the derivatives that are in linear arithmetic representations, computing linear arithmetic representations is converted to compute predicate completely. The improved method is more powerful to generate test data, and can be used for black-box testing.At last, by making up the two algorithms the thesis proposes a framewoek of automated test data generation. And proves the framework is more effieient to derive nonlinear constraints. |
