基于群智能强化学习的电网最优碳-能复合流算法

结合电网能流和碳排放流的传输特性,建立了电网最优碳-能复合流的数学模型,并提出了基于群智能的多步回溯Q(λ)强化学习算法,有效解决了电网碳-能复合流的动态优化问题。其中以线性加权的方式把电网网损、碳流损耗和电压稳定设计为奖励函数,通过引入粒子群的多主体计算,每个主体都有各自的Q值矩阵进行寻优迭代。IEEE118节点仿真结果表明:较传统Q(λ)算法本文所提出算法能在保证较好全局寻优能力的同时,收敛速度至少能提高10倍以上,为解决实际大规模复杂电网的碳-能复合流在线滚动优化提供了一种快速、有效的方法。

Multi-Objective Optimal Carbon-energy Combined-flow of Power Grid Based on Swarm Intelligence Reinforcement Learning Algorithm

Considering the transmission characteristic of carbon emission flow and power flow in power grid , this paper proposes the mathematical model of optimal carbon-energy combined-flow of power grid .Furthermore , this paper a-dopts a PSO-Q(λ) learning algorithm for optimal carbon-energy combined-flow.The carbon emission loss, active power loss and voltage stability are chosen as the optimization objectives on linear weighted way .The algorithm intro-duces multi-agent particle swarm computation , converts the load sections and controllable variables to status and ac-tion, and searches for the optimal action strategy via continuous fault testing , action correction and iteration dynami-cally.Simulation in an IEEE 118-bus system indicates that the PSO-Q(λ) learning algorithm, which improves the convergence speed and maintain the abilities of seeking the global excellent result , providing a feasible and effective way to carbon-energy combined-flow on-line receding horizon optimization in a complex power grid .