ACO-Based Dynamic Routing And Wavelength Assignment Algorithms In ASON

With enduing traditional optical transport network with unprecedented flexibility and expansibility, ASON (Automatically Switched Optical Network) represented the future direction of next generation optical networks. As one of the core technology of ASON, the dynamic routing and wavelength assignment algorithm is the key for ASON to reflecting its intelligence. Undoubtedly, the study on intelligent dynamic RWA algorithm is significant to improve the practicability of ASON. By a perfect RWA algorithm, the state of whole network can be real-timly surveyed, several network resources can be distributed more rationally and the network loads can be balanced high-efficiently. Thus the whole network blocking probability will be reduced effectively and the resource utilization will be improved evidently.In ASON, there is a high demand for the the flexibility of resource scheduling of network. At the same time, the low blocking probability of real-time business should be guaranted. But because of the application of series new technologies, such as wavelength conversion, optical exchange, optical storage, etc. the calculation environment of ASON is further complicated than the traditional optical transport network. So a series of RWA strategy which are designed for traditional optical transport network can not work well in ASON. It has exposed a series of question, such as considering reasonable resource assignment in part-purpose, optimizing goal singly, insufficient flexibility of algorithms, etc.As a heuristic search algorithm, ACA (Ant Colony Algorithm) takes inspiration from the behavior of searching rout from nest to food of real ant colonies. It is used to solve combinatorial optimization problems with its characteristics of positive feedback, distributable and strongly robustness. Considering dynamic routing and wavelength assignment (RWA) problem is in essence a combinatorial optimization problem, this paper designs a new dynamic RWA strategy based on principle of ACA, which is adapted to the environment of ASON networks. The focus of the strategy is more rationally resource assignment in networks with the objective of minimizing connection blocking of the dynamic business.The thesis studies and analyses the relevant technology of dynamic RWA module in ASON, including the routing system of ASON, the influencing factor of dynamic in ASON and the principle of ACA. And a dynamic RWA algorithm model based on ACO (Ant Colony Optimization) in ASON is presented and structured. With above-mentioned relevant technology, a simulation platform is designed to test and improve the algorithm. The main achievements are as follows:(1) The routing technology and system of ASON network are analyzed and summed up, which include multi-area hierarchical routing, distributed wavelength assignment, signaling routing protocol and routing distributing topology, etc. The influence factor and key technology relevant to realize dynamic RWA in ASON are also analyzed in details(2) By studying the principle of AC A, a dynamic RWA algorithm based on ACO is firstly presented. In the algorithm, processing of routing and wavelength assignment sub-problem can be paralleled as a whole. It is able to consider network resource and priority of different business synthetically. Compared with traditional dynamic RWA strategy based on OSPF (Open Shortest Route First), the algorithm can achieve lower blocking probability.(3) The destabilizing factors which influence the execution efficiency in ACA processing are analyzed and summed up. To improve searching efficiency, shorten the process time and prevent the algorithm from sinking into early-maturing trap, the principle of aberrance and local search optimization in algorithm is practiced and information update strategy is adapted. The operation performance of algorithm is tested on simulation platform. Simulation computing showed that the improving measure can improve the performance of algorithm to certain extent.(4) Simulation software is developed to verify the performance of different algorithms and a large amount of test jobs are carried on.(5) The performance of ACA based RWA alrotithm is simulated. According to the result, parameters of algorithm are adapted. And, the algorithm can get better result at the equal resource expenses.