| 摘要: |
| 为求解含复杂约束的多目标优化问题并获得符合实际决策需求的最优解，将多目标优化问题的求解分为2步：1）使用改进的遗传算法对带有较复杂约束的多目标规划模型进行求解，得到Pareto解集；2）基于熵权法构建方案优选评价体系，对Pareto解集进行优选，从而获得多目标优化问题的最佳方案。将本研究方法应用到灌区水资源优化配置问题中检验其可行性与实用性。结果表明：相较于评价函数法获得的结果，Pareto解集可以直观展示不同目标之间相互制衡的关系；根据当地实际情况选取粮经产量比、用水结构信息熵、化肥使用量作为优选指标，优选后的配水方案水分生产力可以达到1.46 kg/m3，总产量达到8.667×107 kg。与传统求解方法比较，本研究提出的求解方法全局寻优能力更强，可以获得更加合理的方案。基于遗传算法与方案优选的多目标优化问题求解方法在求解较为复杂的多目标优化问题时能够获得更为满意的方案，可以为其他多目标问题的求解提供一种新的思路。 |
| 关键词: 多目标求解方法 遗传算法 方案优选 熵权法 水资源优化配置 |
| DOI：10.11841/j.issn.1007-4333.2019.06.18 |
| 分类号: |
| 基金项目:国家自然基金重点项目（51439006） |
|
| A multi-objective optimization model solving method based on genetic algorithm and scheme evaluation |
|
SHAN Baoying, GUO Ping, ZHANG Fan, GUO Shanshan
|
|
College of Water Resource & Civil Engineering, China Agricultural University, Beijing 100083, China
|
| Abstract: |
| In order to obtain the satisfactory solutions of multi-objective programming (MOP) problems,two steps are taken:1) A improved multi-objective genetic algorithm is used to deal with more complex constraints of MOP,and thus the Pareto solution set can be obtained.2) The entropy weight method is used to select a more suitable scheme based on evaluation index system.To prove its practicality and applicability,this method is applied to the water allocation model in Yingke Irrigation District.The results show that the Pareto solution set can fully reflect the relationship between different objectives.According to the practical situation,the ratio of grain to economic,the information entropy of water structure and the amount of fertilizer used are selected as the indicators to value the schemes from Pareto solution set.The water productivity obtained by this method reaches 1.46 kg/m3 and the total yield reaches 86.67×106 kg.Compared with the traditional solution method,the method proposed in this study has stronger optimization ability and can obtain a more reasonable solution.The multi-objective optimization problem solving method based on genetic algorithm and scheme optimization can obtain more satisfactory schemes to solve more complex MOP problems.This method can also provide a new idea to solve other MOP problems. |
| Key words: MOP solving method genetic algorithm schemes evaluation entropy weight method irrigation water allocation |
