求解连续空间优化问题的量子差分混合优化算法

Hybird optimization algorithm based on quantum and differential evolution for continuous space optimization

借鉴量子计算的相关原理和差分进化思想,提出一种用于连续空间优化问题的量子差分混合优化算法。算法的核心是构造由决策向量的分量和量子位概率幅为等位基因的实数编码染色体;采用依据染色体的具体形式设计的互补变异进化部分优秀个体,以加快算法的收敛速度;利用差分进化思想进化部分随机选取个体,以保持算法的全局搜索能力和鲁棒性。对Benchmark函数测试表明,该算法具有寻优能力强、搜索精度高和稳定性好的特点。应用该算法求解路基沉降预测模型参数估计问题,能够有效提高实测沉降数据的拟合精度.

Based on the relational principles of quantum computing and idea of differential evolution, a hybird optimization algorithm based on quantum and differential evolution for solving optimization problems in continuous space is proposed. The core of this algorithm is that, a real-coded chromosome, whose alleles are composed of a component of the decision vector and a pair of probability amplitudes of the corresponding states of a qubit is constructed, and a complementary mutation operator, which is designed based on the specific configura- tion of chromosome, is adopted to evolve some excellent individuals selected to improve the convergence speed of the algorithm, and a differential evolution operator is used to update some individuals selected randomly to keep the global search capability and rubstness of the algorithm. Simulation results on benchmark functions show that the algorithm has the characteristics of more powerful optimizing ability, higher searching precision and better stability. Finally, the algorithm is applied to estimate the paremeters of the prediction model of roadbed settlement, and results show that the prediction model of roadbed settlement based on the algorithm can improve the fitting precision of the observed data efficiently.