摘要
针对寿命终期产品再生决策中较少考虑模块组合效应、模块选择与寿命终期方案分配不同步等不足,提出一种多属性分析与多目标进化算法相结合的模块与方案同步优选方法。兼顾模块寿终质量的影响,建立了以再生收益、模块属性与组合效应最大化为目标的多目标二次优化模型。构建再生属性多准则评价体系,并采用准则交互的模糊层次逼近理想解法将语义评价数据集结为规范化因子融入优化模型中。针对模型的求解,设计一种基于非支配排序的动态灾变进化算法,采用决策空间种群浓度与目标空间进化停滞双重监测下的两级灾变策略保持种群多样性,并实施有限规模跨代竞争。通过对多种报废汽车再生决策的案例分析,验证了所提方法和算法的有效性。
To overcome the deficiencies in recovery decision making of End-Of-Life(EOL)product such as less considering module portfolio effect,non-simultaneous method of module selection and EOL option assignment,a simultaneous optimal selection method for both module and EOL option which combines multi-attribute analysis and multi-objective evolution algorithm was proposed.To achieve the optimal recovery benefits,module attribute and portfolio effect,a multi-objective quadratic optimization model was built by considering the impact of EOL modules quality.A multi-criteria evaluation system of recovery attribute was established and the fuzzy hierarchical TOPSIS with dependent criteria was used to aggregate the linguistic data into normalised indicators to the optimization model.A dynamic cataclysm evolution algorithm based on nondominated sorting was designed as a solution approach.The two-stage cataclysm strategy by monitoring the population density regarding the decision space as well as the stagnation regarding the objective space was used to maintain the population diversity,and the scale of competition across generations was limited.The effectiveness of the proposed method and algorithm was validated in an application of EOL vehicles'decision-making case.
出处
《计算机集成制造系统》
EI
CSCD
北大核心
2015年第2期556-570,共15页
Computer Integrated Manufacturing Systems
基金
国家自然科学基金资助项目(51175262
61105114)
江苏省杰出青年科学基金资助项目(BK201210111)
江苏省科技支撑计划资助项目(BE2014137)~~
关键词
寿命终期决策
产品再生
多属性分析
多目标优化
end-of-life decision making
product recovery
multi-attribute analysis
multi-objective optimization
