Research On The Modeling And Solution Methods Of Disassembly Line Balancing Problem

The serious problems of environmental pollution and resource shortage pose a great threat to human health and social development.Increased rate of advancement in information technology coupled with shorter product life cycles has resulted in a growing amount of premature disposal of products.If these discarded products are not properly disposed of,they will further increase the harm to the environment and cause serious waste of resources.Therefore,industrial recycling and remanufacturing gains vital importance due to the increasing concerns of environmental protection and resource reuse.Product recovery consists of several steps in which the first crucial step is disassembly.Disassembly is defined as the systematic extraction of valuable and hazardous parts from discarded products through a series of operations to use in remanufacturing or recycling after appropriate cleaning and testing operations.In the face of the large-scale product disassembly,the disassembly line being organized in flow shop is the best choice for automated disassembly.The disassembly line balancing problem(DLBP)requires the assignment of disassembly tasks to a set of ordered disassembly workstations while satisfying the disassembly precedence constraints and optimizing the effectiveness of several measures.The research on DLBP has important theoretical value and practical significance.There are four problem studied in this thesis,including stochastic DLBP(SDLBP),DLBP with fixed number of workstations(F-DLBP),twosided DLBP with fixed number of workstations(F-TDLBP),and DLBP with sequencedependent part removal times(SDDLBP).There is a high degree of uncertainty in the structure,quality,reliability,and the condition of the returned products in disassembly.Thus,disassembly task times are assumed random variables with known probability distributions.To cope with the uncertainty related to the disassembly task times,a mathematical formulation for the SDLBP is presented,and a hybrid artificial bee colony algorithm is proposed.A random method is used to generate the initial solutions.A variable neighborhood search method is designed to improve local search efficiency.To effectively choose food sources,a combination selection strategy is proposed for onlooker bees.The scouts explore new food sources based on the best one.The performance of the proposed algorithm is tested on a set of benchmark instances and one case scenario,and the results are compared other approaches.The comparisons demonstrate that the proposed algorithm has superior performance.If the obtained optimal solution is used to organize the disassembly of returned products,disassembly firms can improve the operation efficiency of disassembly lines and remove the harmful parts early to reduce the environmental hazards.To deal with the large-scale returned products disassembled on the disassembly line with fixed number of workstations,an F-DLBP is presented to minimize the cycle time and ensure the balance of workloads among workstations.Then a parallel dynamic neighborhood search(PDNS)algorithm was developed to solve it.In the PDNS,two sets of neighborhood structure were constructed.A dynamic search strategy was applied to realize parallel search of the solutions.The solution can escape from the local optimum rapidly by using the threshold value mechanism.During the adjustment of cycle time,a bound strategy based on binary search is introduced to accelerate the search speed.Finally,the superiority of the proposed algorithm was demonstrated by benchmark instances.The optimal disassembly sequence can effectively shorten the online time of the returned products and increase production of the disassembly lines.Analyzing the operation of large-sized products on two-sided layout disassembly lines,a two-sided DLBP with fixed number of workstations is presented,and a shuffled frog leaping algorithm is proposed to deal with this problem.A combination strategy with multiple heuristics is used to initialize population.The smooth index is used as the fitness evaluation function to divide the population.To speed up evolution speed,the differential evolution algorithm based on variable neighborhood search is used.The elitist strategy is introduced to improve the speed of convergence.Different scale instances are used to verify the feasibility of the proposed model and algorithm.Computational results indicate that the proposed algorithm has superior performance.Through the analysis of disassembly examples,it is further il ustrated that a good disassembly balancing solution can effectively reduce the occurrence of waiting phenomenon in the disassembly process,as well as decrease the idle times,and this will increase the quantity of disassembled products in unit time.In the disassembly process,some parts may interact with each other and their disassembly times will be incremented by additional operations,which will affect the balance of disassembly line.In order to solve the disassembly interference among parts,an SDDLBP is presented with two objectives including minimization of the disassembly cost,and removal of hazardous parts on priority.A Pareto novel artificial bee colony algorithm is proposed.Two heuristic methods are used to construct initial solutions.In order to accelerate the search speed,the onlooker bee phase is abandoned.And some improvements are made in local exploitation and global exploration.A set of benchmark instances is used to test the performance of the proposed algorithm.Computational results evidently indicate the superior efficiency of the proposed algorithm for addressing the environmental and economic concerns while optimizing the multi-objective SDDLBP.The optimal solutions can effectively avoid such situations that the tasks with more time increments are disassembled on priority,which can lower unnecessary workloads and resource consumption.