基于谐振子遗传算法的高效地下水优化管理模型

New Harmonic Oscillator Genetic Algorithm for Efficient Groundwater Optimization and Management

在传统遗传算法和模拟谐振子算法的基础上,结合两者的优点,提出了一种新型快速高效的谐振子遗传算法。通过一个理想的水资源管理模型的算例和一个华北平原典型区地下水资源优化的实际算例,从寻优结果和寻优效率两个方面对谐振子遗传算法、传统遗传算法和模拟谐振子算法进行了对比分析。在两个地下水管理模型中,与传统的遗传算法和模拟谐振子算法相比,新型的谐振子遗传算法搜索效率达到模拟谐振子算法搜索效率的2倍以上,得到的最优解比遗传算法所得到的最优解分别增加供水量1.1×103 m3/d和0.47×108 m3/a,说明谐振子遗传算法具有更强的全局搜索能力和更好的寻优效率。

Combining the advantages of conventional genetic algorithm（GA）with simulated harmonic oscillator algorithm（SHOA）,we put forward a new efficient harmonic oscillator genetic algorithm（HOGA）.Based on the application to an ideal test and an optimization test for the groundwater management model of a typical region in North China Plain（NCP）,the comparisons among GA,SHOA and HOGA are performed in respect both of the quality of optimal solutions and the efficiency of algorithmic optimization.Compared with GA and SHOA,HOGA achieves an efficiency over two times higher than SHOA.The best pumping schemes achieved by HOGA can increase water supply by 1.1×10^3 m^3/d and 0.47×10^8 m^3/a compared with the pumping schemes achieved by GA in the two groundwater management models.This reveals that HOGA has the best global searching abilitywith the highest optimization efficiency.The results suggest that the HOGA presented in this study be of extensive perspectives in application to multi-dimensional and nonlinear groundwater management and optimization.