Fuzzy Logic Based Evaluation of Hybrid Termination Criteria in the Genetic Algorithms for the Wind Farm Layout Design Problem

Abstract

Wind energy has emerged as a potential replacement for fossil fuel-based energy sources. To harness maximum wind energy, a crucial decision in the development of an efficient wind farm is the optimal layout design. This layout defines the specific locations of the turbines within the wind farm. The process of finding the optimal locations of turbines, in the presence of various technical and technological constraints, makes the wind farm layout design problem a complex optimization problem. This problem has traditionally been solved with nature-inspired algorithms with promising results. The performance and convergence of nature-inspired algorithms depend on several parameters, among which the algorithm termination criterion plays a crucial role. Timely convergence is an important aspect of efficient algorithm design because an inefficient algorithm results in wasted computational resources, unwarranted electricity consumption, and hardware stress. This study provides an in-depth analysis of several termination criteria while using the genetic algorithm as a test bench, with its application to the wind farm layout design problem while considering various wind scenarios. The performance of six termination criteria is empirically evaluated with respect to the quality of solutions produced and the execution time involved. Due to the conflicting nature of these two attributes, fuzzy logic-based multi-attribute decision-making is employed in the decision process. Results for the fuzzy decision approach indicate that among the various criteria tested, the criterion Phi achieves an improvement in the range of 2.44% to 32.93% for wind scenario 1. For scenario 2, Best-worst termination criterion performed well compared to the other criteria evaluated, with an improvement in the range of 1.2% to 9.64%. For scenario 3, Hitting bound was the best performer with an improvement of 1.16% to 20.93%.