Local Optimum Network

Local Optima Networks (LONs) represent fitness landscapes as graphs, where nodes are local optima and edges represent transitions between them, providing a powerful tool for analyzing the difficulty of optimization problems. Current research focuses on applying LON analysis to diverse domains, including robot morphology evolution, combinatorial optimization problems (like the Quadratic Assignment Problem and Traveling Thief Problem), and the optimization of neural network architectures. By characterizing the structure of these landscapes through graph properties, LONs offer insights into the performance of metaheuristic algorithms and potentially inform the design of more efficient optimization strategies across various fields.