An adaptive mutation for cartesian genetic programming using an -greedy strategy

The optimization of the number of transistors of combinational logic circuits can lead to faster and cheaper electronic devices, but it is an NP-complete problem. There are deterministic algorithms for this task, but their effectiveness is limited to small problems. Thus, the use of metaheuristics is suitable and Cartesian Genetic Programming (CGP) is widely adopted in the literature. In CGP, mutation is commonly the only operator used for generating new candidate solutions. As a result, the performance of this metaheuristic is dependent on the proper choice of mutation and its parameters. CGP mutation is usually based on uniform mutation and, thus, any modification has the same chance to occur. In order to improve the performance of CGP, a study of the mutation operator is carried out and an adaptive approach using an epsilon-greedy strategy for bias the selection of the node mutation type is proposed here. The proposal is evaluated using a benchmark from the literature. The results obtained indicate that the proposed adaptive mutation is promising, achieving the best mean results in 19 out of the 23 benchmark problems considered here. We also verified a better ability in generating feasible solutions than the standard CGP.