Evolutionary Algorithm And Its Application In Global Optimization And Vehicle Routing Problem

Evolutionary Algorithms (EAs) such as Genetic Algorithms (GAs), Evolution Strategies (ES) and Evolutionary Programming (EP) are a class of stochastic search methods originated from the model of natural evolution. The new EAs are designed and applied to the numerical optimization problems and Vehicle Routing Problems in this thesis. The main contributions are as follows.1. To keep the diversity of the population and avoid the offspring crowding together in some region, a crossover scheme is designed. The offspring generated by this crossover scheme will scatter in different regions. As a result, the diversity of the population can be kept. Based on this, a new evolutionary algorithm is proposed.2. An effective mutation operator is designed, which can make the algorithm escape the local optimum easily. Moreover,a hybrid scheme which can determine whether the proposed mutation operator is used is presented. Based on this, a new evolutionary algorithm is proposed.3. Although the one-point crossover and random mutation operators are easy to implemented, the offspring generated by them for the Vehicle Routing Problems are usually infeasible solutions. In order to make these operators applicable, an improved scheme for these operators is presented. It can always generate feasible offspring. Based on these, a new evolutionary algorithm is proposed for the Vehicle Routing Problems.4. The numerical simulations are made and the results indicate that the three proposed algorithms are effective.