Gregor Molan: Graph based software testing space for software testing optimization

Povzetek. Solution of software testing optimization problem is proposed with introduction of graph based software testing space and three optimization algorithms. The goal is reduction of software testing process costs. The basis is a new definition of testing based on weighted undirected graphs that enables modeling of testing problems with graph theory. Elements of testing process - implementation/development/testing value/status are defined as graph weights. Test suite is defined as a subgraph of test graph and test phase is a set of test suites. Trivial test phase only consists of unit tests. Optimal test phase (OTP) has the maximum implementation value for a given test graph and for a given cost value. The first optimization algorithm, Test Phase Redesign (A1), arranges test phases according to given limitations and eliminates duplicates. The second algorithm, Test Weight Redesign (A2), intervenes in testing amount for a given test phase. The third algorithm, (A3), constructs an OTP. The solution guarantees the maximum software quality according to the given resources.