基于改进多目标粒子群算法的微网双层优化调度策略

Bi-level optimization scheduling strategy of micro-grid based on improved MOPSO

综合考虑微网经济成本、环保成本和系统运行风险程度,建立了多目标优化调度模型,并在迭代末期引入了双层优化,解决了传统优化模型容易漏选最佳解的问题。针对传统多目标粒子群算法的缺陷,提出了基于"栅格-拥挤度"协同筛选策略的多目标粒子群算法。当外部档案中粒子较少时,采用栅格法筛选出全局最优值,当外部档案中粒子较多时,改用拥挤度排序法,从而增强了解集的收敛性和多样性。在下层模型中,建立了基于相对熵组合赋权法的决策算法,综合了主/客观赋权法的优势,使最终结果更加合理。最后以一小型微网为例,验证了考虑双层优化的必要性和改进MOPSO的优越性。

This paper considers economic cost,environmental costs and system operation risk of micro-grid,establishes a multi-objective optimization scheduling model,and introduces a bi-level optimal model at the end of the iteration which solves the problem that the traditional optimization model is easy to miss the best solution. In view of the defects of traditional MOPSO,a MOPSO based on the collaborative filtering strategy of "grid-degree of congestion"is proposed in this paper. When there is few numbers of particles in the external file,we choose the grid method to get the optimum value,when the number is large,we choose the degree of congestion sorting method. As a result,the convergence and diversity of sets are enhanced. The decision algorithm based on the combinatorial weighting method of relative entropy which synthesizes the advantages of methods of subjectively and objectively is established in the lower floor of the method,which makes the final result more reasonable. Finally,the necessity of bi-level optimization and the superiority of the improved MOPSO are verified with a small micro-grid.