【摘要】 软件定义网络（SDN）通过动态调整网络状态以适应工业5.0环境下工业物联网（IIoT）的业务需求，但现有研究在端到端确定性时延的精确建模方面存在不足，且大规模网络中路由与流量调度方法的性能有待提升。对SD-IIoT（软件定义工业物联网）架构进行了研究，基于网络演算（NC）理论构建流量到达与服务曲线的精确模型，同时纳入控制层时延与任务时效性差异的影响；进而建立融合网络资源、业务时延及链路状态的最优化问题，设计编码方式得到基于遗传算法的联合调度策略，通过剪切-拼接方法改进交叉环节，构建可变长度的遗传算法，实现路由与流量的协同优化。实验结果表明，该算法在不同网络规模下均具备优异的确定性时延保障能力与自适应特性。更多还原

【Abstract】 Software-defined networking(SDN) adapts to the business requirements of the industrial internet of things(IIoT) in the Industrial 5. 0 environment by dynamically adjusting the network state. However, existing studies have deficiencies in precisely modeling end-to-end delay and guaranteeing deterministic delay, and the real-time performance of routing and traffic scheduling methods in large-scale networks needs to be improved. The SD-IIoT(software-defined industrial Internet of Things) architecture was proposed. An accurate model of traffic arrival and service curves was constructed based on the theory of network computation(NC), and the influence of control-layer delay and task-timeliness differences was incorporated simultaneously. Furthermore, an optimization problem integrating network resources, service delays, and link states was established, and a joint scheduling strategy based on a genetic algorithm was designed. The crossover operation was improved using a cut-andsplice method to construct a variable-length genetic algorithm, achieving collaborative optimization of routing and traffic. Experimental results show that the proposed algorithm exhibits superior deterministic latency guarantee capability and adaptive characteristics across different network scales.更多还原