MiTS in Depth: An Analysis of Distinct Tabu Search Configurations for Constructing Mixed Covering Arrays

Alan turing work is related with the first use of heuristic algorithms. His work on broking the Nazi code of the Enigma cipher was oriented by a guided search whose expected result in most of the times would be the deciphering of the codes, even though sometimes it might not work. This idea reflects the modern meaning of an heuristic, and represents the main relationship with this chapter, as it involves the use of metaheuristics to try to guide the search to find a solution faster, or a better solution of a problem. The metaheuristic is Tabu Search (TS), and it is used to solve the Mixed Covering Array Problem (MCAP). This problem focuses on the construction of optimal test sets for software testing. The metaheuristic is designed through a fine tuning process that involves the parameters: initialization function, tabu list size, stop criterion, and neighborhood functions. The contributions are: a) a more robust fine tune process to design a new TS approach; b) the analys is of parameter values of the TS; and, c) new bounds over a benchmark reported in the literature.