Multiobjective Particle Swarm Algorithm With Fuzzy Clustering for Electrical Power Dispatch
Multiobjective Particle Swarm Algorithm With Fuzzy Clustering for Electrical Power Dispatch
复制标题
DOI：
10.1109/tevc.2007.913121
复制
发表时间：
2008-10-01
影响因子：
14.3
通讯作者：
Tiwari, Manoj Kumar
中科院分区：
文献类型：
--
作者：
Agrawal, Shubham;Panigrahi, B. K.;Tiwari, Manoj Kumar
经济调度是一个高度受限的优化问题，包含决策变量之间的相互作用。由于化石燃料发电机的运行而引起的环境问题将经典问题转化为多目标环境/经济调度（EED）。本文提出了一种基于模糊聚类的粒子群（FCPSO）算法来解决涉及冲突目标的高度约束的 EED 问题。 FCPSO 使用外部存储库来保存搜索过程中发现的非支配粒子。所提出的模糊聚类技术在不破坏帕累托前沿特征的情况下将存储库的大小管理在限制范围内。结合了 Niching 机制，将粒子引导至帕累托前沿较少探索的区域。为了避免陷入局部最优并增强粒子的探索能力，提出了一种自适应变异算子。此外，该算法结合了基于模糊的反馈机制，并迭代地使用信息来确定折衷解决方案。该算法的性能已在标准 IEEE 30 总线六发电机测试系统上进行了检验，生成了均匀分布的 Pareto 前沿，其最优性已通过针对 epsilon 约束方法的基准测试得到了验证。结果还表明，所提出的方法获得了高质量的解决方案，并且能够在几乎所有试验中提供令人满意的折衷解决方案，从而验证了所提出的方法对现实世界多目标优化问题的有效性和适用性。
Economic dispatch is a highly constrained optimization problem encompassing interaction among decision variables. Environmental concerns that arise due to the operation of fossil fuel fired electric generators, transforms the classical problem into multiobjective environmental/economic dispatch (EED). In this paper, a fuzzy clustering-based particle swarm (FCPSO) algorithm has been proposed to solve the highly constrained EED problem involving conflicting objectives. FCPSO uses an external repository to preserve nondominated particles found along the search process. The proposed fuzzy clustering technique, manages the size of the repository within limits without destroying the characteristics of the Pareto front. Niching mechanism has been incorporated to direct the particles towards lesser explored regions of the Pareto front. To avoid entrapment into local optima and enhance the exploratory capability of the particles, a self-adaptive mutation operator has been proposed. In addition, the algorithm incorporates a fuzzy-based feedback mechanism and iteratively uses the information to determine the compromise solution. The algorithm's performance has been examined over the standard IEEE 30 bus six-generator test system, whereby it generated a uniformly distributed Pareto front whose optimality has been authenticated by benchmarking against the epsilon-constraint method. Results also revealed that the proposed approach obtained high-quality solutions and was able to provide a satisfactory compromise solution in almost all the trials, thereby validating the efficacy and applicability of the proposed approach over the real-world multiobjective optimization problems.