多能流系统合作协同的不确定多目标决策

Cooperative co-evolutionary-based uncertain multi-objective decision making for multi-energy flow systems

中文关键词  多能流系统  不确定性  多目标优化  合作协同进化  决策制定  平衡控制

英文关键词  multi-energy flow systems  uncertain  multi-objective optimization  cooperative co-evolutionary  decision making  balance control

中文摘要

      多能流工业生产过程具有多目标、强耦合、时变、不确定性等特点, 针对此类系统的平衡调度问题, 本文提 出一种基于合作协同优化的不确定多目标决策方法. 以钢铁企业副产煤气系统为例, 针对系统未来状态的不确定 性, 本文在优化决策的过程中结合卡尔曼滤波方法和贝叶斯定理, 提出一种考虑条件预期的不确定决策模型. 该模 型能够同时分析当前目标和预期目标, 从而消除未来状态不确定性带来的影响. 针对副产煤气系统多能流强耦合的 特点, 本文在优化决策过程中综合考虑单能流系统特性以及多能流系统的协同关系, 基于图模型原理提出基于双向 权重的协同进化方法, 从“总体”–“局部”相结合的角度给出最优的决策策略. 通过实际钢铁企业数据的仿真实验 表明, 该方法能够充分考虑未来的不确定性, 同时兼顾单能流系统性能和多能流耦合关系, 给出合理的调度决策方 案. 该方法可用于具有多目标、强耦合以及不确定性的复杂多能流系统, 为其调度决策问题提供支持.

英文摘要

      Aiming at the balance scheduling problem of the multi-energy industrial production system, which has the characteristics like multi-objective, strong coupling, time-varying, uncertainty, etc., a cooperative co-evolutionary-based uncertain multi-objective decision making method is proposed. Take the byproduct gas system of iron and steel enterprises for example, considering the uncertainty of the future state, the optimal scheduling decision making strategy is given under the consideration of the maximal objective and the expected one based on the anticipate flexible multi-objective decision making method by incorporating the Kalman filtering method and Bayes theorem into the optimizing process, which eliminates the influence of the uncertainty in the future. Besides, in order to solve the strong coupling problem among multi-energy flow when making the optimal scheduling strategy, comprehensive considering the characteristics of the single energy flow system and the cooperative relationship of the multi-energy flow system in the decision-making optimization process, a bi-directional weight based cooperative co-evolutionary method which combines with the graph model principle is proposed to make the decision strategy from the“global”–“local”coordinated angle. The simulation experimental results by using the industrial data demonstrated that the proposed method could give the optimal scheduling decision making strategy with fully considering the future uncertainty, the single energy flow property and strong coupling characteristic among the multi-energy flow. Thus, the proposed method could be used to provide support for the scheduling decision making problem of the complex systems with multiple targets, strong coupling and uncertainty.