A relational multi-objective genetic algorithm

We propose a new relational multi-objective genetic algorithm (RMOGA) with a novel generic operator, inheritance. It is found that inheritance have very good performance when used with a multi-objective genetic algorithm. Its characteristic is similar to a crossover operator. However, it aims to exchange the mathematical relationships, but not values, between two selected sub-chromosomes. We also propose to use a Pareto-ranking to help form a fitness function. Then it is optimized by the relational multi-objective genetic algorithm, which includes the new operator. The advantage of this approach is its ability to inherit the best relationships. To test the effectiveness of the proposed methodology, experiments on using it for assembling broken objects were performed. Excellent results are obtained.