面向业务可靠承载的电力弹性光网络自主协同决策

Autonomous Collaborative Decision-Making for Power Elastic Optical Network Oriented to Service Reliable Bearing

针对电力通信骨干网面临频谱资源受限、调配灵活度差等挑战,提出一种面向业务可靠承载的电力弹性光网络自主协同决策方法。首先,建立长时业务可靠性约束,并通过虚拟队列漂移实现队列可靠性约束偏差感知;然后,将电力弹性光网络中业务与资源协同问题建模为业务请求、路径、频隙之间的三维匹配问题,并通过路径与频隙的聚合使其降维为一对一匹配问题;最后,引入升价匹配算法解决多业务请求对网络资源的竞争。仿真结果表明,相较于基于最小化队列漂移加惩罚的路由与频隙分配(MDPP-RSA)算法和基于首尾精确适配的路由与频隙分配(FLEF-RSA)算法,该文所提算法可有效提升电力业务可靠承载能力和网络资源利用率,平均业务队列积压降低了51.87%、80.18%,频谱利用率提升了33.52%、54.41%,并可满足不同优先级业务的高可靠传输需求。

The power communication backbone network is an important infrastructure to ensure the safe and stable operation of the power system by means of power grid service reliable bearing.However,cross domain,layer and level source-grid-load-storage efficient interaction and multi energy wide interconnection also bring challenges to the service data bearing capacity of the power communication backbone network,where the network is facing increasingly prominent problems such as limited spectrum resources and poor deployment flexibility.Aiming at such problems,this paper proposes an autonomous collaborative decision-making method for power elastic optical network(EON)oriented to service reliable bearing,which integrates new generation information communication technology to upgrade traditional technology applications and provide support for wide area transmission and regulation of electric energy.First,he long-term reliability constraint of different types of power grid services and the optimization problem in the power EON are constructed,optimization goal of which is maximizing weighted network utility considering service priority,spectrum utilization and spectrum fragmentation.Furthermore,due to the randomness of power grid services and the time-varying nature of power EON resources,Lyapunov optimization is introduced to transform the long-term optimization problem and constraints into a series of short-term deterministic optimization problems, where the service reliability constraints deviation awareness is realized through virtual queue drift without any prior statistics and future prediction information. Finally, the coordination between service and resource in power EON is modeled as a three-dimensional matching problem among service requests, paths and frequency slots, which dimension was reduced to a one-to-one matching problem through the aggregation of path and frequency slot. Considering the competition of multi service requests for network resources, the price-based matching algorithm is utilized to achieve the autonomous collaboration between them in a unified time period, so as to ensure the reliable bearing of power services by the power communication network. The simulation results show that compared with the min drift plus penalty-based routing and spectrum allocation algorithm and the first-last exact fit-based routing and spectrum allocation algorithm, the proposed algorithm has better performance in network utility, spectrum utilization improvement, and service reliability guarantee. Specifically, the average network utility increased by 18.45% and 35.71%, the spectrum utilization increased by 33.52% and 54.41%, the cumulative throughput increased by 9.86% and 15.17%, the service request queue backlog decreased by 51.87% and 80.18%, and the virtual queue backlog decreased by 54.50% and 81.45%, which can also meet the high reliable transmission requirements of different priority services. As prospective theoretical research on the integration of electric power and information communication, the research results of this paper can provide a feasible technical scheme for the deployment of the next generation power communication backbone network, improve the service data carrying capacity of the power communication network, and provide a guarantee for the development of large-scale emerging power communication services. In the future research, based on the highly time-varying nature of service and network under the background of new power system construction, the autonomous collaborative decision-making method with lower complexity and stronger adaptability will be further studied.