Research On Methods For Recovery Decision-making And Disassembly Planning Of End-of-life Products Based On Ubiquitous Information

End-of-life(EOL)product recovery is gaining attention due to increasing pressures related to resources consumption and environmental pollution.To ensure an effective and profitable product recovery business,recovery decision-making and disassembly planning are two critical issues.Recently,the development of ubiquitous information technologies(e.g.Internet of Things,intelligent sensors,RFID)have a potential to enable efficient EOL data management by information tracking and sharing through the product lifecycle and closed supply chain.However,how to identify the optimal decision for recovery management remains as a challenge.There is still a lack of a comprehensive optimization framework for solving this problem.The purpose of this paper is to fill this gap.A series of optimization models and approaches were proposed to offer theoretical contribution and support for facilitating more efficient,flexible and intelligent recovery decision-making and management.The following issues were studied in this paper through mathematical modeling,algorithm simulation,numerical experiments and applied research.First,quality-driven recovery decision-making for EOL components based on ubiquitous information was studied.Two assessment methods were developed for component group and component individual,respectively.For the former,a fuzzy logic model was utilized to make full use of the product lifecycle information to obtain a comprehensive quality index.For the latter,a new model was proposed by combining conditional distribution estimation of remaining useful life(RUL)with recovery cost-benefit analysis,where conditional RUL distribution was predicted by utilizing both the failure data and condition monitoring data based on a proportional hazard model.Further,RUL was considered as a quality measurement to enable a quality-driven decision model.Under RUL uncertainty,an interval decision-making approach was developed to identify a satisfactory solution according to decision maker's risk preferences.Second,this paper studied the problem of performance evaluation and optimized selection for recoverable EOL products in the context of ubiquitous information.Two new problems were presented and solved.One was how to select a set of EOL products under constraints to achieve optimized solution;the other one was how to identify the optimal recovery portfolio during multiple EOL modules.A multi-phase hybrid decision model was designed to provide a general decision framework for the first problem.The proposed hybrid model considered multiple and conflicting goals and practical constraints by using fuzzy cost-benefit analysis,multi-criteria evaluation and goal programming.The second problem was formulated as a quadratic optimization model.Both the quality and correlation effect of alternative EOL modules were considered in this model.Then a combined method of multi-criteria interaction analysis and multi-objective optimization was used to solve it.Moreover,a dynamic catastrophic genetic algorithm was developed to seek for the optimal recovery portfolio and options.Third,an integration model of the recovery option selection and disassembly planning for EOL product remanufacturing and dismantling was proposed and solved.A quality-driven profit maximization model was formulated for this integrated problem.The interaction of recovery option selection and disassembly planning was considered.In order to determine recovery options,disassembly level,and disassembly sequence simultaneously,this paper developed an improved co-evolutionary algorithm(ICA)to search for an optimal EOL solution.An integrated method of multi-target reverse recursion and partial topological sorting was proposed to generate a feasible EOL solution that also reduced the complexity of genetic constraints handling.The proposed algorithm adopted the evolutionary mechanism of localized interaction and endosymbiotic competition.Further,an advanced local search operator and global disturbance strategy were introduced to improve the convergence performance.Computational experiments under various scenarios were conducted to validate the superiority of the developed ICA.Further,the paper tracked the problem of trade-off between product remanufacturing and dismantling from a perspective of sustainability.A multi-objective optimization model with no demand constraints was first proposed to identify the optimal product recovery solution that best balanced the economic,environmental and societal performances for sustainability based on the availability of ubiquitous information.The Pareto optimal solutions were obtained by using Non-Dominated Sorting Genetic Algorithm II.Further,a new decision model considering both EOL quality and second-market demand was developed.Then,a dual population co-evolutionary algorithm was designed to search for the optimal recovery strategy,options,yield,and disassembly solution.A case of automobile engine was studied to test and validate the effectiveness of the two models.Moreover,this paper investigated how the quality,cost,and price impact the recovery strategy and its sustainable performance based on the first model.The impact of the demand on the sustainable performance and optimal recovery yield was also analyzed and discussed by scenario simulation.Finally,two real cases of EOL product(ammeter and lifter)were investigated to further demonstrate the potential applications of the proposed models and approaches.An integrated decision model for multi-tier products was proposed to identify the optimal sequence,level and recovery options for EOL ammeter disassembly.The impacts of various costs and requirements on the recovery profit and solution were also discussed.A tactical recovery decision-making problem was investigated for the used lifter.The results showed that the proposed model and approach could provide a strong support for EOL product management.