Research On Multi-Objective Optimization Model And Intelligent Algorithm Of OTN Networking

With the rapid development of mobile communication technology,the traffic of the backbone network grows fast and data transmission volume is hugely increasing due to the large amount of services.Optical Transport Network(OTN)technology which inherits the benefits of synchronous digital system and wavelength division multiplexing technologies has such advantages as the electrical domain and wavelength/sub-wavelength multiplexing and cross-connect,flexible exchange granularity and carrying more types of customer signals.These features help improve the networking efficiency and controllability,and also provide more reliable network protection and recovery capabilities,which can solve the problems of ultra-large capacity and ultra-long-distance transmission for network operators.During the process of OTN network operation,the complex scenario considering the multi-objective resource optimization problems has become one of the current research hotspots.In the complex and practical network scenarios where there exist multiple physical lines between directly connected nodes,shared risk link groups(SRLG)and associated services groups,the optical network resource optimization algorithms need to provide a set of resources allocation schemes whose multiple objectives are non-dominated to each other in order to meet the diverse network requirements.Based on this,this thesis mainly relies on the key special sub-project of the National Key Research and Development Program of China in "Broadband Communication and Novel Networks"-Research on the application of novel 100G/200 G optical networks with engineering demonstration(2019YFB1803605),and studies the modeling of multi(many)-objective resource optimization problems in OTN networking and its multi(many)-objective intelligent solution algorithms.This thesis firstly introduces the optical network and its research background.The second chapter involves the research basis of this thesis,including the overview of OTN and its survivability,the research status of multi-objective resource optimization model and multi(many)-objective evolutionary algorithms in OTN,so as to establish the foundation for the coming work of this thesis.The research contents and contributions of this thesis are as follows:1)The thesis constructs a multi-objective resource optimization model for the OTN networking and studies the corresponding multi-objective intelligent algorithm.Currently,most of the existing studies are solved OTN resource optimization problems by the single-objective integer linear programming or heuristic algorithms,and the research schemes are rarely considering the numerous constraints in practical applications.Therefore,based on the actual networking application scenarios,this thesis designs a three-objective optimization model for OTN networking and optimizes the total length of services carrying paths,the load difference of the primary links and the overall unreliability of a network.To solve this model,a Multi-objective Hybrid Evolution with Bi-indicators Awareness(MHEBA)algorithm is proposed,which preprocesses services information according to the mutually exclusive nature of the service requests and designs a unique individual encoding strategy based on the preprocessed service requests information.In addition,both the bi-indicators awareness and the hybrid evolution mechanism are specially designed to adaptively select the evolution mode,which increases the convergence of the algorithm and improves solutions’ distribution.Finally,the algorithm performance simulation is completed in the network topologies of Internet2 and Cernet,and the effectiveness of the proposed MHEBA algorithm is verified by both the comparison of simulation results and the comprehensive evaluation of multi-objective evaluation indicators.2)Combined with the survivability of OTN,the many-objective optimization model in OTN networking is built with five diversionary objectives: the total length of service carrying paths,the load difference of the primary links,the overall unreliability of a network,the number of repeaters used in a network and the standard deviation of the number of services carried by the SRLG.To solve this model,a wavelength weighting strategy with the minimum number of repeaters is designed to facilitate the services to select the appropriate wavelength,which reduces the number of repeaters required additionally.Moreover,a Many-objective with Multi-factor Hybrid Awareness(MMHA)algorithm is further proposed,and a population diversity maintenance strategy along with a relaxed domination strategy and a dimension reduction strategy are designed to optimize the solutions and display performance of the proposed algorithm.Considering the above factors,the performance simulation of the proposed MMHA algorithm in Cernet topology verifies that the MMHA algorithm can obtain more solutions with superior convergence and better diversity from the perspectives of simulation results and multi-objective evaluation indicators,thus illustrating the effectiveness of the MMHA algorithm.