摘要

Membrane computing is a new branch of natural computing, whose aim is to abstract computing ideas from the structure and the functioning of living cells to construct computing models and algorithms. The obtained algorithms are generally called membrane-inspired algorithms or membrane algorithms, which are known as a class of intelligent algorithms inspired by biological behaviors of living cells. In this work, a membrane algorithm with a three-levels of hierarchical cell-like structure is proposed, where elementary membranes can dynamically evolve to generating working space (performing cell division) during the computation. We test our method by solving 14 instances of a benchmark of the vehicle routing problem, as well as a numerical example of solid waste management in Chengdu, China. Experimental results show that our method performs well on solving the 14 instances of the benchmark of the vehicle routing problem comparing with the genetic algorithm, tabu search algorithm, simulated annealing algorithm, ant system algorithm, and evolutionary algorithm. In practice, our method can quickly obtain the global optimal solution of the numerical example of solid waste management.