Optimization of Multilevel Hierarchical Transshipment System in Logistics Transportation Based on Fish-Swarm Algorithm

Guangqiu Huang,Yao Yuxia,Yan Ren

2007-01-01

Abstract:A large-scale nonlinear optimum programming model to solve the optimization problem of multilevel hierarchical transshipment system in logistical transportation was established.In the optimization model,the solving space was discretized based on actual consideration of space constraints of variables and precision of solutions;the policy of discretization was to create grids with certain size within space of variables,and the optimal locations for all transfer stations at each level of a multilevel hierarchical transshipment system were selected from all known coordinate points.Because the dimension of the model was enormous,an improved fish-swarm algorithm was used to solve the problem.In the improved algorithm,two matrices describing the relation between transfer stations of all levels with angular points of grids and the relation between two adjacent levels' transfer stations were established to delete constraints and compress greatly variables of the model;an evolutionary algorithm based on evolutionary similarity was applied to mix the basic properties of global and local search;the self-adaptive delta variation operator,dual arithmetic crossover operator,peak-jump operator were used to improve the behaviors of artificial fish-swarm algorithm.An application result shows that the speed and reliability of the optimization algorithm is improved greatly.