| Product configuration is a key technology in mass customization production.Different from the research of general product configuration solution technology,we focus on the product configuration phase,the procurement phase,and the assembly phase to help customized companies complete the production of customized products from a holistic view.At the same time,it is important to evaluate the various uncertain factors in the production of enterprises to reduce the losses caused by uncertainty.Therefore,the main research contents of this article are as follows:First,we consider the joint optimization of product configuration and supply chain,while considering the impact of carbon quotas and carbon tax regulations that have been implemented in most countries and regions on product configuration.A two-stage method combining PSO and greedy algorithm is proposed to solve product configuration problems with configuration rules and carbon emissions problems.For the two-stage method,PSO is employed in the first stage to find a feasible configuration composed of alternative module instances,and in the second stage,a greedy algorithm is applied to obtain the best purchase amount of the corresponding module instances in the configuration.An example of a configurable driller is presented to prove the effectiveness of the proposed method.Also,the different impact of carbon cap and carbon tax on the allocation results are compared.Numerical experiments are performed to verify the effectiveness of the configuration model and the efficiency of the proposed algorithm.In addition,the model also extends the single sourcing strategy,and analyses the different influence of single sourcing networks and multiple sourcing networks on product configuration.Second,considering the assembly process,we develop the joint optimization model of integrated product configuration,supply chain decision-making and assembly task scheduling.In the assembly process,the influence of different module instances on three different assembly structures was creatively considered.In addition,we implement the overall carbon emission optimization,and a multi-objective optimization model is established.To solve the model,an enhanced ε constraint method is designed.By modifying the objective function and adding artificial variables,the enhanced ε-constraint method avoids searching for too many weak Pareto solutions,and improves the search efficiency.The calculation results of the model show that: total cost and carbon emissions are a pair of conflicting goals;decision-making preferences and delivery dates have an important impact on corporate decision-making.In order to further verify the effect of the algorithm,we also complete a large-scale numerical experiment to compare the enhanced ε-constraint method and the Weighted Sum Method to verify the effectiveness of the algorithm.Finally,we focus on the product configuration problem under uncertain environment.Using box uncertainty set and ellipsoid uncertainty set,the uncertainty of purchase price and assembly time is described.After transforming the model with uncertain parameters into a robust counterpart that is easy to solve,through a driller case,we evaluate the influence of parameter uncertainty(perturbation ratio)and decision-making risk preference(safety parameter)on the product configuration model,and the constraints in the model are calculated probability bound indicates the feasibility of the robust solution.At the same time,we analyse the impact of the delivery date on the configuration result,and calculate the upper limit of the constraint violation probability in the model and the corresponding configuration cost to verify the reliability of the robust solution.In order to further study the applicability of robust optimization to product configuration problems,numerical experiments have been completed in this paper.The experiment shows that,in the two robust models,the total cost of the configuration results of the ellipsoidal robust model solution is the smallest and the robustness is better.In addition,decision makers can also set safety parameter based on risk bias to minimize the total cost. |
