Research On Optimization Models And Algorithms For The Problem Of Resource Allocation In Elastic Optical Networks

In recent years,the booming of internet based applications,such as internet protocol television,video on demand,cloud computing,requires an efficient networking infrastructure.Wavelength division multiplexing(WDM)networks cannot suit for various bandwidth connection requirements adaptively because they can only provide the coarse granularity of a single wavelength.The recent elastic optical networks(EONs)can provide the flexible and variable bandwidth allocation to each connection request and get higher spectrum utilization by using the technology of orthogonal frequency division multiplexing(OFDM)with multicarrier modulation technology.It can distribute the high-speed data stream into several orthogonal low-speed subcarriers and make the network more flexible.In spite of these advantages,EONs also have to face many challenges,such as routing and spectrum assignment,virtual optical network mapping in EONs,datacenter placement and VNF deployment.So,the investigation of EONs have became a hot topic.The main work and contributions of this dissertation are as follows:(1)Routing and spectrum assignment problem in EONs is investigated.First,when the security of the nodes in network is considered,a global constrained optimization model,which minimizes the maximum index of used frequency slots,is established.To solve the established model,we design a genetic algorithm with well designed crossover,mutation and local search operators.Second,different from previous works,the hierarchical property in the process of determining the scheme of routing and spectrum assignment is considered.We also establish a bi-level optimization model with the energy consumption and the maximum index of used frequency slots as the leader's and follower's objectives to be minimized,respectively.A hybrid genetic algorithm framework is proposed to solve the model by integrating a heuristic algorithm into a genetic algorithm.Simulation experiments are conducted,and the experimental results indicate the effectiveness of the proposed model and algorithm.(2)Virtual optical network mapping in EONs is investigated.First,virtual optical network mapping in EONs with multi-core fibers is investigated.we establish a constrained optimization model to determine the optimal scheme of optical network mapping,core allocation and spectrum assignment.To solve the model efficiently,a tailor-made encoding scheme,crossover and mutation operators are designed.Based on these,an efficient genetic algorithm is proposed to obtain the optimal schemes of the virtual nodes mapping,virtual links mapping and core allocation.Second,to tackle the problem of virtual optical network mapping in EONs with single-core fiber.The hierarchical property in the process of virtual nodes mapping and virtual links mapping is considered.We establish a bi-level programming model with the energy consumption of the optical networks and the maximum index of used frequency slots as the leader's and follower's objectives to be minimized,respectively,which are used to determine the optimal schemes of virtual optical nodes mapping,virtual links routing and spectrum assignments.Simulation experiments are conducted,and the experimental results indicate the effectiveness of the proposed model and algorithm.(3)The problem of resource allocation for VNF service chains(VNF-SCs)in inter-DC EONs is investigated.First,each datacenter can only provide some specific VNFs.The resources of both bandwidth and datacenter are considered.We divide VNFs into two sets:one contains all independent VNFs and another contains all dependent ones.To solve this challenging problem,we establish a global optimization model for this problem.An efficient genetic algorithm with a tailor-made encoding scheme is proposed to solve the model.The simulation experiments are conducted on several situations,and the results indicate that the proposed model is reasonable and the proposed algorithm is efficient.Second,we investigate a network planning problem,which requires to determines the optimal scheme of datacenter placement(including the number and location of datacenters),and the optimal schemes of routing and VNF deployment for VNF-SCs.We establish a bi-level optimization model,where leader's objective is to minimize the number of datacenters and make the VNF deployment as balanced on datacenters as possible in order to determine the datacenters number and locations,and follower's objective is to minimize the maximum index of used frequency slots and the number of used frequency slots.To solve the whole model effectively,a tailor-made crossover,mutation and local search operators are designed.Based on these,an efficient memetic algorithm is proposed.Simulation experiments are conducted,and the experimental results indicate the effectiveness of the proposed model and algorithm.