基于试验遗传算法的平原圩区除涝排水系统最优规划

平原圩区除涝排水系统最优规划是一个多维、非线性复杂系统优化问题。本文提出基于试验优化设计思想的试验遗传算法，即在标准遗传操作中增加了均匀设计、正态随机和摄动调优等试验操作技术。某圩区排水系统最优规划实例表明，与可分规划法及简单遗传算法相比，试验遗传算法能严格满足系统约束条件，且操作简单，计算效率更高。由于该算法兼顾了整体空间的种群多样性和局部空间极小值的搜索效率，同时能根据需要自动调整计算精度，因此较易获得复杂问题的全局最优解。

Optimization of planning for drainage system in plain polder area based on experimental genetic algorithm

optimization of planning for drainage system in plain polder area is a complicated multi-dimensional non-linear optimizing problem. In this paper, the experimental genetic algorithm based on the experimental optimization design is proposed to derive global optimized solution of complicated problem. This algorithm is based on the standard genetic algorithm but experimental operation, techniques such as uniform design, normal random distribution and variables perturbation are embedded. The comparisons of the calculation results of an example show that comparing with the dividable programming method and simple genetic algorithm, the proposed method is more strictly subject to the limited space of variables and more easy to be used and having better optimizing efficiency. Because of both the individual variety of whole space and the efficiency of searching for local minimum point are taken into account and the calculation precision can be automatically self-adjusted, the proposed method possesses the ability to obtain the best global solution.