Combinatorial Test Generation for Multiple Input Models with Shared Parameters.

Area of Science:

• Software Engineering
• Computer Science
• Testing and Verification

Background:

• Combinatorial testing commonly uses a single input model to generate test sets for t-way coverage.
• Handling multiple input models with shared parameters presents challenges in test set generation and redundancy reduction.

Purpose of the Study:

• To address the problem of combinatorial test generation for multiple input models with shared parameters.
• To propose an efficient approach for generating multiple test sets that satisfy t-way coverage across all models while minimizing redundancy.

Main Methods:

• Formal definition of the combinatorial test generation problem for multiple input models.
• Development of an efficient algorithm to generate multiple, optimized test sets.
• Experimental evaluation on five real-world applications.

Main Results:

• The proposed approach significantly reduces redundancy between test sets for multiple input models.
• The method demonstrates superior performance compared to post-optimization techniques.
• Achieved t-way coverage across all input models with minimized overlap.

Conclusions:

• The developed approach is effective for combinatorial test generation with multiple input models and shared parameters.
• This method offers a practical solution for reducing test suite size and improving efficiency in complex testing scenarios.