A novel improved ant colony algorithm is proposed, which is aiming at the locally optimal solution and highly complex issues in solving the logistics distribution model of E-commerce. First, corresponding problem description of logistics distribution procedure of E-commerce was analyzed, dynamic travelling salesman problem model was established with taking minimizing logistics time as objective function, and appropriate constraints were set. Second, framework of the improved ant colony algorithm was raised through integrating advantages of both local search algorithm (LSA) and ant colony algorithm. All operations like constructing solutions, pheromone update policy and local search in the novel algorithm were setting. Time complexity reduced to O(n~4) in theoretically. Last, offline performance representing algorithm criterion was simulated based on analog data. Results show that, offline performance of the novel algorithm improves obviously, which would provide a reference to apply it in real logistics distribution scene in future.
展开▼
