Improvement Of Ant Colony Algorithm And Application Research In TSP And VRP

Ant Colony Algorithm is a novel swarm intelligent algorithm by studying the behavior of natural ants looking for food,Its main characteristics are positive feedback,parallelism and being easy to combine with other algorithms.It attracts close attention of the scholars at home and abroad from its coming and was rapidly applied to solve many combinatorial optimization problems achieved good results.Ant Colony Algorithm tends to produce the local optimum and stagnation phenomenon because it depends on random selection strategies,phenomone updateing,and positive feedback of optimum strength to search much better solution. In order to improve the convergence rate on the basis of solution quality,this thesis firstly summaries the current research state and application,basic principles,mathematical model and parameter setting,analyses the reasons from defacts of typical algorithm AS and ACS.For the defacts of slow convergence and prematurity, phenomone initialization and updating and stimulating factor were improved.A new improvement algorithm DDS-ACS has been proposed in this thesis based on diverse initial phenomone strategy,dynamic updateing of local and global phenomone and strength of the stimulating factor. DDS-ACS has been successfully applied to solve the Vehicle Routing Problem(VRP) and logistics after further optimization according to their individual characteristics.The results show that the DDS-ACS is much better convergence and gobal search capability than AS, ACS and other mentioned papers in this thesis,at the same time,which proves DDS-ACS has better feasibility and effectiveness some extent.