Beherrschung rasant wachsender Variantenvielfalt im Testmanagement von varianten- und versionsreichen mechatronischen Automatisierungsprodukten

Manufacturers of mechatronic products have to be able to adapt to more individual customer requirements and the increasing flexibility of production plants in the context of Industry 4.0 [1]. This requires a drastically increasing flexibility of the components contained in the production plants, which leads to an increasing number of variants and versions of their products. At the same time quality assurance measures become more complex. The increasing number of variants entails an increasing number of necessary test cases. However, the time available for testing is usually limited so that only a selected number of test cases can be executed. This selection is still made manually and based on the experience of the responsible tester. Due to the increasing number of variants, the test cases that should be executed based on occurred changes are difficult to identify for the tester. Hence, there is the danger that already evaluated test cases will be repeated and that critical test cases are missed [2]. This inefficient testing practice is reflected directly in the quality of the products and causes a loss of time and money in the company. This article presents a method that automatically selects and prioritises test cases out of a large amount of test cases for multi-variant products. In order to have a better overview of all product variants and versions of a manufacturer, an approach similar to feature modelling is presented and evaluated prototypically. Using the feature model of a product family as well as already existing and tested variants, test cases are systematically selected with minimal effort. Hereby, the resource-optimized test planning uses the similarity of products of a product line. The individual variants are similar in terms of the installed feature and the requirements to the product variant. Based on the proposed approach, test cases that test the differences between the variants are selected automatically, as the differences are still untested in contrast to the similarities between the variants. Hence, only the changed parts are tested. In addition, the effects of these changes on other features of the product is considered for the test case selection. If only a limited amount of time is available for testing, the tester has to prioritise the selected test cases so that the most important ones are executed. The method presented in this paper performs this prioritisation automatically. For this purpose, an algorithm is used which gathers, weighs and calculates the priority of the test cases based on properties such as their criticality, execution time and historic data. As a result, the algorithm provides a recommendation on the execution order of the test cases. Hence, valuable test resources can be saved through automated test case selection and prioritisation. Further, an improved quality assurance, which is independent of the tester’s experience, is guaranteed. The proposed concepts are evaluated on several industrial examples in the development of mechatronic components and show a significant reduction of time and costs within the companies.