Cycle Mutation: Evolving Permutations via Cycle Induction

Evolutionary algorithms solve problems by simulating the evolution of a population candidate solutions. We focus on evolving permutations for ordering such as traveling salesperson problem (TSP), well assignment quadratic (QAP) and largest common subgraph (LCS). propose cycle mutation, new mutation operator whose inspiration is well-known crossover operator, concept permutation cycle. use fitness landscape analysis to explore characteristics which works best. As prerequisite, we develop distance measures: distance, k-cycle edit distance. The predicts that better suited mapping than it problems. experimentally validate these findings showing mutation’s strengths QAP LCS, its limitations TSP, while also less prone local optima commonly used alternatives. integrate into open source Chips-n-Salsa library, metrics JavaPermutationTools library.