Generating test-cases from an object-oriented model with an artifical-intelligence planning system

Abstract

Black-box test-generation requires a model of the system under test to describe what is to be tested. Testing criteria and test objectives define how it is to be tested. This paper describes an approach to black-box test-generation in which an AI (artificial intelligence) planner is used to generate test cases from test objectives derived from UML (Unified Modeling Language) Class Diagrams. The UML Class Diagrams are conceptual models of the systems under test. They differ from traditional design and requirements models in that they include information pertinent to test case generation. From these models, test objectives and a domain theory are: obtained, transformed to planner representations, and input to the planner. The planner uses the problem description to generate a test suite that satisfies the UML-derived test objectives. This paper describes the application of the testing approach to an industrial problem.