Genetic operators for the atomic cluster problem

This paper investigates the recent advances in applying genetic algorithms to non-bonded molecular conformation problems. Through the use of deterministic local optimisation to reduce the search domain and innovative phenotype genetic crossover operators, genetic algorithms are now able to efficiently solve problems such as determining the minimum energy configurations of pure atomic clusters ( N =2, …,100), mixed atomic clusters ( N = 2,…,20) and molecular clusters ( N = 2,…,32). A comparison of the more conventional genotype crossover operators with the more recently developed phenotype operators is presented.