一种动态改变权重因子和交叉率的混沌差异演化算法

Chaos differential evolution algorithm with dynamically changing weighting factor and crossover factor

针对非凸、多峰及非线性函数优化难的问题，在差异演化（DE）算法的基础上进行了改进，提出了一种动态改变权重因子和交叉率的混沌差异演化算法（CDDE）．在该算法中引入进化速度因子和聚集度因子．权重因子被表示为聚集度因子的函数，交叉率被表示为进化速度因子的函数．在每次迭代时利用聚集度因子动态改变权重因子，利用进化速度因子动态改变交叉率，从而使算法具有动态自适应性．引入混度变异算子，利用混沌的随机性和遍历性使算法具有较强的跳出局部最优值的能力．通过对3个典型函数的测试表明，CDDE算法收敛速度明显好于DE，收敛精度得到了提高．

Based on Differential dynamically changing weighting Evolution （DE）, a new differential evolution algorithm with chaos mutation factor and factor and crossover factor （CDDE） is presented to solve the optimization problem of the non-convex, multimodal and nonlinear function. The evolution speed factor and aggregation degree factor of the population are introduced in this new algorithm. The weighting factor and the crossover factor are formulated as functions of the aggregation degree factor and the evolution speed factor respectively. In each iteration process, the weighting factor is changed dynamically based on the current aggregation degree factor, and the crossover is changed dynamically based on the current evolution speed factor. The chaos mutation factor is introduced to avoid trapping into the local optimum. DE and CDDE are tested with three well-known benchmark functions. The experiments show that the convergence speed of CDDE is significantly superior to DE, the convergence accuracy is also increased.