改进飞蛾火焰算法在多目标水资源优化配置中的应用

The Application of Improved Moth Flame Algorithm to Optimal Allocation of Multi-objective Water Resources

水资源紧缺局面愈演愈烈,迫切需要进行水资源优化配置,以此提高水资源利用效率。基于自然元启发的优化算法越来越多的应用于求解多目标水资源优化配置问题。飞蛾火焰算法全局收敛性能较差且容易陷入局部最优解,针对以上问题,结合帕累托最优解概念,引入快速非支配排序策略,利用拥挤度及拥挤度比较算子对飞蛾火焰算法进行改进,继而采用改进的飞蛾火焰算法对ZTD1、ZTD2及ZDT3等3个多目标函数进行仿真实验,结果表明计算得到的帕累托最优前沿接近于理想帕累托前沿,收敛性能及精度优于本文所列举大部分多目标优化算法。最后将改进飞蛾火焰算法应用于三亚市水资源优化配置中,得到了23组帕累托最优解。选择了对缺水量最小有特殊偏好的方案作为最终决策方案,结果显示三亚市不同用水户总需水量39015万m^3,各用水户总分配水量为39015万m^3,缺水量为0,产生的经济效益为203.91亿元。改进的飞蛾火焰算法为多目标水资源优化配置问题提供了新的求解方法。

The shortage of water resources is intensifying, and there is an urgent need for the optimal allocation of water resources to improve the utilization efficiency of water resources.Nature-inspired meta-heuristic optimization algorithms are increasingly applied to solve multi-objective optimal allocation of water resources.Considering that the moth flame algorithm (MFO) has poor global convergence performance and is easy to fall into the local optimal solution, we adopt the concept of Pareto optimal solution, introduce a fast non-dominated sorting strategy, the crowding distance and the crowded-comparison operator on the MFO.Then the improved moth flame algorithm is used to simulate three multi-objective functions such as ZTD1, ZTD2 and ZDT3.The results show that the obtained Pareto optimal front is close to the true Pareto optimal front, and the convergence performance and accuracy are better than most of the multi-objective optimization algorithms listed in this paper.Finally, the improved moth flame algorithm is applied to the optimal allocation of water resources in Sanya City, and 23 Pareto optimal solutions have been obtained.The scheme with the special preference for the minimum water shortage is chosen as the final decision- making scheme.The results show that the total water demand of different water users in Sanya City is 39 015×10^4 m^3, the total water allocation of each water user is 39 015×10^4 m^3, the water shortage is 0, and the economic benefit is 203.91×10^8 RMB.The improved moth flame algorithm provides a new method for multi-objective water resources optimization configuration problem.