Software testing is important in any software project to detect faults. Many test case design techniques offer various methods to produce effective test cases. Test suite size is the biggest concern of test design technique to overcome the exhaustive testing problem due to time and resources. As most of the faults are caused by interactions of few parameters, t-way testing is an appropriate technique to be used. In this work, t-way testing taxonomy and existing strategies that support t-way testing has been explored. The existing strategies have been compared against factors in the taxonomy. Although beneficial, the existing strategies tend to support a specific type of t-way support interactions either uniform strength, variable strength or Input-Output based Relations (IOR). Those types of support interactions function in different situations perfectly. However, software tester received testing problems in various situations. Therefore, a single flexible strategy that allows software tester to choose the most suitable support interactions based on testing problems on hand is required. In addition to that, t-way test suite generation has fallen under the NP-hard problem. There is no single strategy that can guarantee the optimal results at all times. As a consequence, T-way Test Suite Generation based on Ant Colony algorithm (TTSGA) was proposed in this research. Besides supporting all three types of support interactions, this strategy applies metaheuristic search technique, which is never been applied by any other strategies that support all types of support interactions. Metaheuristic search technique also is proven to produce minimum test suite size as compared to computational search technique. As the name suggested, the strategy adopts the ant colony algorithm for the search technique. Performance of the TTSGA strategy was implemented using several benchmarked experiments for each type of support interactions. Besides that, two statistical analysis tests, Friedman Test and Wilcoxon Rank test has been performed. The results have been compared to other metaheuristic and computational strategies for all types of support interactions. As far as test suite size is concerned, TTSGA achieved the first rank for variable strength and IOR uniform configurations with the mean rank of Friedman Test is 1.49 and 2.00 respectively. The strategy also produces encouraging results for uniform strength (2.75 in third rank) and IOR non-uniform configurations (3.25 in second rank). Generally, TTSGA produces good results in high interactions strength and configurations.
