Research On Combinatorial Testing Technology For Boolean-Specifications

Combinatorial testing is an important technical branch in software testing and has been widely used in industry.Although the combinatorial testing technology has been widely studied,there are still some problems that cannot be solved well: on the one hand,people are concerned about whether the combinatorial testing technology has sufficient error detection capabilities in different scenarios;on the other hand,people also care about Accurately describe the characteristics of software input fields that can cause errors.At the same time,the Boolean-specification testing is also an important branch of software testing,because the execution path of a software program usually depends on the judgment result of the Boolean expression in the predicate statement being true or false.Therefore,this article will try to study the combinatorial testing related technologies for Boolean-specifications.This article focuses on two aspects.On the one hand,the fault detection capability of the combinatorial testing technology in Boolean testing is compared with the random test technology as a benchmark.The ability to describe the characteristics of an incorrect input field.The experimental objects selected in this paper are 20 Boolean expressions extracted from the TCAS system,and19131 non-equivalent mutation expressions are generated by 10 mutation operators.The specific research work carried out in this paper includes the following two aspects:(1)Select 20 Boolean expressions and their variations from the TCAS system as experimental objects,and compare the fault detection capability of the combinatorial testing with the fault detection capability of the random test.The experimental data show that the combinatorial testing has better error detection ability than the random test.In addition,through the specific analysis,the following rules were found: when the strength of the combinatorial testing is relatively high,the combinatorial testing's ability to detect faults is significantly stronger than that of the random test;when the software input domain dimension is high or the fault FTFI value is large,the combinatorial testing and random testing have similar fault detection capabilities.(2)Using 20 boolean expressions and their variations extracted from the TCAS system as experimental objects,compare the description capabilities of the minimal failure-causing schema and the probabilistic failure-causing schema to the characteristics of the input domain that can cause fault.Experimental data shows that probabilistic failure-causing schema is significantly better than the minimal failure-causing schema in terms of recall rate,accuracy,and comprehensive evaluationindicators,which indicates that the probabilistic failure-causing schema is significantly stronger than the minimal failure-causing schema in describing the characteristics of the input domain that can cause fault.