一种基于混沌序列的动态差分进化算法在电力系统动态经济调度中的应用

针对电力系统动态经济调度(DED)问题，引入差分进化算法，提出一种基于混沌序列的动态差分进化算法(ＡＤＤＥＣＳ)。该算法采用混沌序列动态调整差分进化算法的参数设置，保持种群的多样性。动态搜索策略被用于提高算法的整体搜索性能，它由全局搜索策略和局部搜索策略2部分组成。为了加速收敛和解决DED复杂的约束处理问题，采用基于多目标概念的约束处理机制，并提出一种根据机组调节能力来按比例分摊不可行解约束违反量的新方法。同时在搜索过程中，通过采用不同的变异策略结合改进的随机搜索策略来避免算法早熟,增强全局最优解的搜索能力。提出的方法的可行性和有效性由10机测试系统来证明，和其他方法相比，ADDECS方法计算速度快，计算精度高且鲁棒性强。

Dynamic Differential Evolution Algorithm Based on Chaotic Sequences for Dynamic Economic Dispatch Problem of Power System

Differential evolution (DE) algorithm is applied to solve dynamic economic dispatch (DED) problem and a dynamic differential evolution based on chaotic sequences (ADDECS) is proposed. Chaotic sequences is used in dynamic parameter settings adjustment in DE to keep population diversity. Dynamic search strategies including both global and local search strategy are used to improve algorithm efficiency. To accelerate convergence and deal with complicated constraints of DED problem, constrained process mechanism based on multi-objective concepts is adopted. A new constraint process approach is proposed to allocate infeasible constraint violation proportionally among units according to their regulatory abilities. At the same time, different mutation strategies and improved random search are used to guide population evolution and to enhance global search ability. The feasibility and effectiveness of proposed method is verified on a 10-unit test system. Compared with other methods, ADDECS can obtain better solutions in shorter computational time along with higher robustness.