Enhancing Path-Oriented Test Data Generation Using Adaptive Random Testing Techniques

TL;DR

This paper introduces an adaptive random testing approach for path-oriented test data generation, improving efficiency by reducing invalid inputs and effectively handling complex constraints through dynamic domain partitioning.

Contribution

It presents a novel divide-and-conquer method that enhances path coverage testing by efficiently generating valid test data with fewer rejects.

Findings

Reduces invalid test inputs compared to existing methods.

Effectively handles complex path constraints.

Demonstrates practical benefits through experimental results.

Abstract

In this paper, we have developed an approach to generate test data for path coverage based testing. The main challenge of this kind testing lies in its ability to build efficiently such a test suite in order to minimize the number of rejects. We address this problem with a novel divide-and-conquer approach based on adaptive random testing strategy. Our approach takes as input the constraints of an executable path and computes a tight over-approximation of their associated sub-domain by using a dynamic domain partitioning approach. We implemented this approach and got experimental results that show the practical benefits compared to existing approaches. Our method generates less invalid inputs and is capable of obtaining the sub-domain of many complex constraints.