Research On The Modeling And Optimization Of Disassembly Sequence Planning Problem For Used Products

With the development of society,enterprises manufacture more and more products to meet people's growing demand and consumption.A tremendous amount of energy and rescources are being consumed during the processes of manufacturing,use,and retirement while having a great impact on the environment.How to reduce the resource consumption,rational use of limited resources,and achieve sustainable development of human society are the matters of common concern to all countries.It is well known that the products will be obsoleted once they are end-of-life(EOL).However,these EOL products,such as electronics,vehicles,and industrial equipment,often contain valuable components and materials that can be reused or recycled,value recovery.By the recovering the residual value embodied in EOL products,it helps reduce large virgin material and energy consumption and thus contributes to the sustainable development.Disassembling the used products including obsoleted electronics,vehicles,and industrial equipment,etc,is the first step to recycling and reusing them,which is an important study for researchers that focus on the recovery of EOL products.It has received increasing attention in recent years.During the disassembly process,obsoleted products are separated into subassemblies including components and parts.The selection of disassembly sequence significantly affects the disassembly efficiency and recovered value,so the recycling and reuse of obsoleted products rely heavily on the disassembly plan.A disassembly sequence planning(DSP)has arisen to find the best order of product disassembly for maximum recovery value and disassembly efficiency.In this paper,we focus on the modeling and optimization of DSP from different perspectives so as to fill the gap in the current research,which are presented as follow:(1)We study a selective sequential DSP considering sequence-dependent costs among disassembly operations.A mathematical model is proposed with the objective to maximize the recovery profit using an AND/OR graph(AOG)subject to sequence-dependent costs.In the AOG,‘AND' and ‘OR' relations illustrate the relationships between subassemblies.Moreover,an improved genetic algorithm(IGA)is proposed to solve the problem.The performance of the proposed IGA is measured on a series of test instances against exact approaches including CPLEX and an iterative method.(2)We introduce energy efficiency to the DSP and present a multi-objective selective sequential DSP,whereby not only the profit is maximized,but also energy consumption is accounted as an important indicator.A multi-objective artificial bee colony(MOABC)algorithm is proposed to solve the DSP for the Pareto front.To demonstrate the efficiency of MOABC,we apply it to two different scales of instances and further compare with the non-dominated sorting genetic algorithm(NSGA-II).(3)We study a synchronous parallel DSP(SPDSP)problem with more than one operator,which provides an effective way to disassemble large or complex products.Two optimization objectives i.e.disassembly time and profit,are considered in the SPDSP.According to the characteristitics of the problem,a disassembly hierarchical tree graph(DHTG)is built to represent parallel disassembly tasks while a DHTG-based multi-objective selective SPDSP is modeled.A MOABC is adapted to different scales and complexities of instances.The proposed method is also compared with NSGA-II to demonstrate its effectiveness.(4)SPDSP simplifies the modeling and complexity of parallel DSP,but it reduces the efficiency of parallel disassembly at the same time,which imposes that all operators must keep synchronous work.Thus,this paper proposes an asynchronous parallel DSP(APDSP)model.An IGA is developed to solve the APDSP problem.The experimental results show that APDSP outperforms traditional SPDSP in terms of the disassembly efficiency.In one word,this work aims at modeling DSPs considering the representation of disassembling products,evaluation indicators(i.e.,optimization objectives),and operational mode(i.e.,the operators' configuration),while developing population-based evolutionary algorithms to optimize the proposed DSPs and obtain the high-quality disassembly solutions.The study of DSPs in this paper provides more alternatives for adapting the complexity and the flexibility of disassembly practices in the future.