A novel evolutionary algorithm for multi-constrained path selection

Qi Xiaogang,Liu Lifang,Liu Sanyang

DOI: https://doi.org/10.1007/11941439_74

2006-01-01

Abstract:For the problem of multi-constrained path selection, a novel evolutionary algorithm named MCP_EA is proposed. Firstly, a novel coding technology named PNNC (Preceding Natural Number Coding, PNNC) is designed, and no circle exists on the path coded by PNNC. Secondly, a novel crossover operator called DCC operator (Dispersing Connection Crossover operator, DCC operator) is designed to guarantee the validity of the crossed paths and the diversity of the population. Thirdly, a novel mutation operator named selective mutation operator is proposed. Finally, the theoretical analysis proves that the algorithm converges to the satisfactory solution with probability 1.0. Extensive simulations show that the novel evolutionary algorithm outperforms the H_MCOP in performance for the problem, and is a promising algorithm for the problem with high performance.