Abstract
This pilot study forms part of research into the optimization of the shape of a wave energy collector to improve energy extraction using genetic algorithms. Two main types of genetic algorithms exist, differentiated by the use of binary or real numbers as object descriptors. The study is intended to ascertain if one type is more suited to the specific problem by comparing the performance of two example algorithms. The algorithms optimize the shape of a bisymmetric wave energy collector moving in two degrees of freedom (surge and pitch). The collector is described by ruled surfaces in one quadrant, defined by the positions of seven vertices. The cost function is based upon a first-order model of the system, with the collector optimally tuned to a number of incident regular waves with a generalized occurrence distribution. High velocities and large collector volumes are penalized. An assessment of the performance of the two algorithms is made, looking at the improvement in value and change in diversity of the respective populations. A comparison is also made of the computational requirements of the different parts of the optimization process.