The mathematics of software testing using genetic algorithm

Abstract

A Markov Chain approach to estimate reliability of a software system using genetic algorithm is presented. In this approach, the code is initially converted into control flow graph and then reduced to a dd-graph. The fitness function of the genetic algorithm is calculated based on the path coverage. The edges of the dd-graph are assigned weights based on the Markov transition probability matrix and the value of the fitness function is calculated as the sum of weights of all edges of the dd-graph covered by the test suite. In this paper, arithmetic crossover and insertion mutation is applied for the floating point populations and genetic algorithm is used to generate test cases to achieve path coverage with less computation cost and time. The proposed approach results in identifying the most critical path of the system.

Evolutionary algorithms for path coverage test data generation and optimization: a review

“…One example is shown in Figure 1, which shows the different steps carried out to find the linearly independent paths for a specific program. With the help of path testing, the test cases are created and executed for all possible paths which results in 100% statement coverage and 100% branch coverage [21][22].…”