Research On Optimal Allocation Method Of Manufacturing Resources In Cloud Manufacturing

In order to integrate social manufacturing resources,improve resource utilization and reduce manufacturing costs effectively and respond to the market demand quickly,a new web-based and service-oriented intelligent manufacturing model called cloud manufacturing was proposed by Academician Li Bohu.The optimal allocation problem of manufacturing resources in cloud manufacturing is one of the core issues of cloud manufacturing,and the performance of its solution has great influence on the operation quality and efficiency of cloud manufacturing service,and then affects the development and promotion of cloud manufacturing.Therefore,the optimal allocation problem of manufacturing resources in cloud manufacturing is intensively studied in this thesis.The research contents and outcomes of this thesis are as follows:(1)Basic theories about optimal allocation of cloud manufacturing resources are explored.Operational elements,related notions,operational models and the architecture of cloud manufacturing are discussed in detail.Then the general process for allocating cloud manufacturing resources is presented,and features and complexity of problems with the cloud manufacturing resources allocation are pointed out.Besides,the key processes stages for optimal allocation of cloud manufacturing resources are illustrated.(2)In light of the modelling and mapping problems of cloud manufacturing services and cloud manufacturing resources,ontology-based modelling methods and mapping methods based on extensible clustering algorithms are proposed.First of all,cloud manufacturing services are divided into eight categories,including Design as a Service,Simulation as a Service,Test as a Service.Fabrication as a Service,etc.Secondly,cloud manufacturing resources are classified into basic information,functional information,evaluation information and state information in light of their attributes according to needs of cloud manufacturing for corresponding resources.Thirdly,mathematical expressions based on extensible basic-element models are put forward for models of cloud manufacturing services and cloud manufacturing resources.Fourthly,mapping methods are proposed for cloud manufacturing services and cloud manufacturing resources based on extensible clustering algorithms by integrating correlation functions of extenics with hierarchical clustering.At last.the mapping method is verified to be feasible and effective by example analysis.(3)To evaluate cloud manufacturing resources,extensible comprehensive evaluation method is put forward based on extension superiority and Fuzzy Analytical Hierarchy Process.At first,an index system covering Quality of Service and Capability of Service is put forward for evaluating cloud manufacturing resources.Subsequently,the extensible comprehensive evaluation method based on extension superiority is investigated according to evaluation characteristics of cloud manufacturing resources,and a comprehensive method for-evaluating cloud manufacturing resources is put forward by combining extensible comprehensive evaluation method with Fuzzy Analytical Hierarchy Process.At last,the evaluation method is verified to be feasible and effective by example analysis.(4)Problems with optimal allocation of cloud manufacturing resources combination are divided into two categories,including the problems with demand preference and without demand preference.Thus,methods based on bi-level programming and non-cooperative game are put forward for optimal allocation of cloud manufacturing resources combination.Firstly,an index system involving Quality of Service and Flexible index is proposed for optimal allocation of cloud manufacturing resources combination with different structures.Secondly,a model based on bi-level programming is put forward for the optimal allocation of cloud manufacturing resources combination with demand preference,and a solution based on Fast and Elitist Non-dominated Sorting Genetic Algorithm is proposed for solving the model.For the optimal allocation of cloud manufacturing resources combination without demand preference,a model based on non-cooperative game is proposed and solved by improved Fast and Elitist Non-dominated Sorting Genetic Algorithm.At last,feasibility and effectiveness of both methods are verified by example analysis.(5)It is proposed that parameters of above models can be estimated based on Artificial Bee Colony algorithm.Firstly,the problem for acquiring parameters of above models is converted into parameter estimation for models for system identification.Secondly,an optimal model is built for optimal allocation of cloud manufacturing resources combination for the purpose of minimizing loss functions of estimated values.Meanwhile,a solution is put forward to solve the model based on Artificial Bee Colony algorithm.Finally,the feasibility and effectiveness of the model and the solution are verified by example analysis.