多目标模糊评价遗传算法综合质量交换网络

Mass Exchange Networks Synthesis Using Multi-Objective Genetic Algorithm Based on the Fuzzy Reasoning

综合质量交换网络时, 评价指标往往不止一个, 多个指标相互之间往往又是无法比较、相互矛盾的,甚至有的指标只是定性的,无法定量地表示。故尝试应用模糊数学对多目标进行评价,有效地模拟了多目标决策的过程, 过程中规则的变换可反应决策者对不同目标的重要性取向。此外还提出了多目标模糊评价遗传算法(MOGAFR)求解多目标优化问题。数例的求解证明了该法可以得到非劣解集合,并可以对非劣解集中的个体区分出优劣,规则不同, 可以收敛到非劣解集合的不同部分,多目标质量交换网络的综合证明其解决工程问题的有效性。所提出的方法适用于求解其他系统的多目标优化问题。

When synthesizing Mass Exchange Networks, there are usually several objectives. The objectives are generally incomparable and incompatible. Some objectives are qualitative and can not be quantitatively expressed. In this article we evaluated multiple objectives using fuzzy mathematic method which can effectively simulate the process of reasoning and decision with multi-objectives. The Multi-Objective Genetic Algorithm based on the Fuzzy Reasoning(MOGAFR) proposed in this paper can optimize the multi-objective problem. A mathematic example showed that MOGAFR can get the Pareto optimal set and identify the most suitable solutions for the rules in the Pareto optimal set. It can also get different part of Pareto optimal set by changing rules in the process. A case study for a Multi-Objective Mass Exchange Networks synthesis problem showed that MOGAFR can solve the engineering problem effectively. MOGAFR can also solve the multi-objective problems of other systems.