| High-end equipment manufacturing(HEM)is a strategic equipment manufacturing which determines the comprehensive competitiveness of a country.It has the characteristics of leading the development of high technology,being at the core of the industrial chain and occupying the top of the value chain.In order to enhance the core competitiveness of the industry and seize the commanding heights of future technology and economic development,countries are actively introducing the new generation of manufacturing paradigm(NMP)based on cloud computing and Internet of Things into HEM to realize the digital,grid and intelligent transformation of HEM.The manufacturing service Composition and Optimization selection(MSCOS)is the core technology of the NMP,which integrates emerging big data technology with combinatorial optimization theory to improve the manufacturing efficiency and quality of service(Qo S).However,the recent explosion of pandemic and geopolitical uncertainties has had a significant impact on MSCOS of HEM.How to explore the influence mechanism of uncertain manufacturing environment(UME)on MSCOS,and establish the resilience management mechanism of MSCOS to resist the impact of UME under the framework of MSCOS has become a new challenge.This paper combines the MSCOS in the new generation manufacturing paradigm with the classic resilience management theory,and carries out relevant research from two major directions: the design of the generalized MSCOS resilience framework and the solution of the practical problems of typical high-end equipment MSCOS.Firstly,the distributed system architecture and service-oriented software architecture of the NMP are studied,and the architecture for HEM resilience management is proposed.Then,the resilience management scheme and service resilience assessment scheme for HEM are proposed.Finally,aiming at the practical problems of MSCOS for three types of HEM,including complex technology,large-scale heavy equipment and military equipment,solutions are proposed based on the three steps of generalized MSCOS.The main research contents and contributions are as follows:(1)A generalized MSCOS resilience framework was constructed to meet the requirements of resilience management for HEM in UME.Firstly,a service system architecture based on cloud computing,fog computing and edge computing is constructed to meet the requirements of resilience management such as resource allocation management and demand load balancing.Secondly,a software architecture based on microservices is established to meet the requirements of system fault tolerance,lightweight overhead and multi-layer coordination of services.Then,after discussing the resilience management theories in different fields,a resilience management scheme based on big data analysis,service resilience evaluation and capacity replacement is proposed.Finally,an evaluation scheme of manufacturing service resilience was proposed based on the recovery time model of supply chain resilience management theory.(2)Aiming at the problems of numerous participating manufacturing services with widespread redundancy in HEM with complex process type,an MSCOS method based on precise service matching was proposed.Firstly,combined with knowledge base of HEM field,hierarchical task network is adopted to decompose tasks into task chain containing multiple sub-tasks.Secondly,by fuzzifying manufacturing service Qo S data,a triangular fuzzy number similarity matching algorithm for redundant services is proposed,which can match candidate services for each subtask and filter redundant services.Then,taking MSCOS in aerospace engine manufacturing as the research scenario,a multi-objective optimization model was proposed with production cost,manufacturing time,service resilience and supplier output as objective functions.Finally,an adaptive simulated binary crossover(SBX)operator based on relaxation factor and penalty variable is proposed considering the large-scale variation of optimization problems.The adaptive SBX operator replaces the fixed distribution factor with a nonlinear function.The adaptive SBX operator is applied to NSGA-III to solve the combinatorial optimization problem.(3)Aiming at the problems of logistics expensive,high-energy consumption and difficulty in punctuality in HEM with large-scale heavy-duty type,a MSCOS method for task chain reconfiguration is proposed.Firstly,the driving scenario for the reconfiguration of the traditional task chain is described,and the reasons for the low efficiency of logistics transportation caused by feasible but inefficient solutions generated by the multi-granularity attribute of manufacturing service are explained,and the crossgranularity reconfiguration algorithm of the task chain is proposed.Then,a multiobjective optimization model was proposed with production cost,manufacturing time,service resilience and energy consumption as objective functions.Finally,a dual track NSGA-III is proposed.In this algorithm,a clustering-based judgment mechanism of offspring population diversity was added,through which a suitable environmental selection mechanism was selected from two environmental selection mechanisms to retain offspring elite individuals.Then,the convergence of optimal solution is improved and the integrity of pareto front is guaranteed.(4)Aiming at the problem that manufacturing time of military equipment is difficult to accurately control,a MSCOS method based on Qo S time attribute lean evaluation is proposed.Firstly,the software architecture based on microservices is expounded as the premise of Qo S time attribute lean evaluation.Secondly,based on the cooperative scheduling theory,a method is proposed to leanly evaluate the Qo S time attributes of manufacturing services through the supplier’s human resource and material resource information.Then,a multi-objective optimization model with production cost,manufacturing time,service resilience and service reliability as objective functions is proposed based on MSCOS problem of a military unmanned aerial vehicle(UAV).Finally,an improved NSGA-III based on Chebyshev metric is proposed.The algorithm selects offspring elite individuals through hyperplane projection clustering and intracluster Chebyshev metric sorting,so as to improve convergence of Pareto optimal solution and diversity of solution space.The problem of MSCOS in HEM under UME is not only the actual demand of manufacturing enterprise development,but also an important direction of MSCOS research field.The established MSCOS resilience framework integrates the HEM resilience management throughout the generalized MSCOS process,and can provide theoretical support for the collaborative resilience management of the whole process of MSCOS.The proposed system architecture solves the problems in HEM such as untimely decision making,delayed information feedback and underlying control force.The proposed software architecture has the characteristics of fault tolerance,multi-layer service coordination,and lightweight overhead,which provides containerized application deployment for HEM MSCOS technology.The proposed resilience management scheme for HEM and service resilience evaluation scheme provides a theoretical basis for the resilience management of UME.The three solutions and combinatorial optimization algorithms proposed for three different types of HEM enrich the domain knowledge of MSCOS technology in HEM and improve the adaptability,convergence and diversity of combinatorial optimization algorithms,which have certain theoretical significance and practical value... |
PDF Full Text Request