Time is On The Side of Grammatical Evolution

The computational complexity of Evolutionary Algorithms (EAs) is a well-known concern. This paper is concerned with the resource consumption of GELAB, a novel Grammatical Evolution (GE) system implemented in Matlab. GE is an evolutionary technique for program search that manipulates large populations of computer programs over multiple generations. In this paper, we present our reflections on some recently conducted experiments with GELAB. GELAB leverages the power of Matlab to perform hybrid optimization on its individuals. The results show that we can substantially decrease computation time without compromising the accuracy of the system.