Neuro-fuzzy–based optimization techniques
Asit Kumar Nayek, Jibendu Kumar Mantri, Arpan Adhikary, Sunanda Hazra · 2025
Soft computing techniques are pillar stones of intelligent systems when applied to real-life engineering problems like proportional integral derivative (PID) controller design, finding maximum power transfer point, controller for hybrid power distribution grids, power generation forecasting and usage, maximizing power generating efficiency, dynamic load management and many more. These problems are mapped into fuzzy systems to generate fuzzy membership functions and are fed into neural networks to adopt the parametric variabilities of an electrical system. In neuro-fuzzy hybrid systems, optimization plays a vital role. There are some essential characteristics of metaheuristic-heuristic supervised algorithms that fall under the population-based nature-inspired algorithm category. All these features are termed adaptive neuro-fuzzy inference system (ANFIS), and they escalate the performance of optimization. In this chapter, we discuss some of the ANFIS techniques that are popularly used in engineering applications. Out of the recently used methods, particle swarm optimization (PSO), craziness-based particle swarm optimization (CRPSO), recursive least square (RLS) algorithm, neuro-fuzzy convex optimization, etc. are discussed. Comparisons are discussed after reviewing their performance in different segments. The coherence between deep neural network (DNN) and fuzzy-based systems is discussed in this context.