Multi-Objective Optimization in Genetic Algorithms

Abstract

Genetic algorithms (GA) are a fascinating optimization technique designed to explore vast spaces, mimicking the way genetic material works in living organisms. Even a small group of individual examples can effectively navigate a large area because they carry schemata—these handy substructures that can be mixed and matched to create stronger individuals. Research into competing schemata reveals that the most effective way to replicate them is to boost their numbers exponentially based on their relative fitness. This is exactly the approach that genetic algorithms take. To determine fitness, a wide range of individual fitness cases is examined. This process can be quite efficient, especially if the fitness cases themselves evolve using their own genetic algorithms. In the first chapter of the Introduction to GA, various search methods are discussed in detail, including Hill Climbing and Simulated Annealing. The biological background of GAs is also covered broadly, touching on DNA structure, methodology, coding, and the types of parameters involved, all illustrated with a simple GA program. The book explores optimization methods, emphasizing two primary types: multi-objective optimization and single-objective optimization. There’s a thorough discussion on both, showcasing the best search techniques for each type of optimization.