The effect of GA redundancy on the design of reverse logistics network

Genetic algorithm's approach has been one of the most important optimization techniques though the related Problems of selection pressure still exist. Diversity is the key issue. For this reason, we explored the implication of redundancy in both genetic algorithm's overrepresentation and multi-explanation. Our research result supports the view of crosscompetition but not diversity loss when considering uniform coding redundancy with expressive elitism and the tolerance of infeasible solutions. On the other side, we developed new decoding procedure based on the inborn searchability of the small world. In designing the network of the reverse logistics having multiple lower-dimensional constraints, with the help of multiple overlapped fitness functions, the performance of genetic algorithms can truly be improved.