| Computer software system has a profound impact on human society. It increasingly highlight-s the importance of software testing. Reducing the cost and improving the efficiency of software testing has a important practical significance and economic value. This paper investigates on t-wo aspects of software testing system from the perspective of discrete event dynamic systems-software testing technology and software testing workflow. It discusses about improving the efficiency of software testing by the following ways:increasing the degree of test automation, ana-lyzing workflow performance’s bottlenecks and allocating human resources rationally, optimizing task scheduling and execution priorities.Firstly, this paper proposes an algorithm that can automatically generate test cases based on Petri net model. The algorithm does a thorough search of all states of software system to fully cover any expected state that software operations may arise and to eliminate the omission of possible software errors due to a limited test coverage of manually-written test cases. The feasibility of the algorithm is demonstrated by the software prototype developed by PyQt and an example of the course purchase feature from NetEase Study.Secondly, in this paper we modelled the software testing workflow using generalized stochas-tic Petri nets, analyzed the model with simulation software, acquired the characteristic of time and resource from various parts of workflow and figured out bottlenecks that affect system per-formance. Through model’s equivalent simplification, we can clearly see each part’s time perfor-mance impact on the overall system, which is beneficial while guiding the allocation of resources in industry software testing.Furthermore, this paper proposes four scheduling schema and three selecting schema after the in-depth analysis of the part "test case manually execution" which restricts the software testing performance, and the research on testing task scheduling and priority. Thus we got twelve com-binations from the four scheduling schema and the three selecting schema. Then we developed a simulation program of task scheduling based on Petri net, compared the performance of each scheduling option in six different situations and made the analysis of their differences. |
