On the Performance of BRKGA in the Multi-objective Offshore Wind Farm Layout Optimization Problem

The placement of wind turbines in offshore farms requires simultaneously minimizing construction costs and maximizing energy production, giving rise to the Multi-objective Offshore Wind Farm Layout Optimization Problem (MoWFLOP). As an NP-hard problem, MoWFLOP has been widely addressed using metaheuristic approaches. However, the literature does not report results based on the Biased Random-Key Genetic Algorithm (BRKGA) for this problem. This paper investigates the performance of two BRKGA variants for the MoWFLOP: BRKGA/D and NS-BRKGA. The former is based on the decomposition technique of MOEA/D, whereas the latter relies on the non-dominated sorting mechanism of NSGA-II. Experiments on 100 benchmark instances show that BRKGA/D outperforms NS-BRKGA in terms of solution quality and diversity. However, both variants are outperformed by MOEA/D and NSGA-II. These findings highlight the limitations of BRKGA strategies and motivate further research into extensions for multi-objective optimization.