Research On Resource Service Composition Optimal-selection Problem In Cloud Manufacturing

Cloud manufacturing is a new service-oriented, high efficiency and low consumption, knowledge based, networked, and agile manufacturing paradigm, which develops from but extends the traditional manufacturing modes and systems, and integrates many advanced technologies such as cloud computing, the Internet of things, virtualization, servitization, high performance computing, artificial intelligence technologies, and so on. And all manufacturing resources and capabilities in the product lifecycle are virtualized and then encapsulated as cloud services. Resource service composition(RSC) means the currently existing resource services are integrated into composite services to be invoked for addressing a manufacturing task that a single service cannot do. And selecting the optimal RSC to successfully execute a manufacturing task is one of the key issues to achieve the resource or service value-add in cloud manufacturing. The problem of RSC optimal-selection is a typical NP-hard combinatorial optimization problem with the characteristics of multiextrema, multiobjectives, nonlinearity and uncertainty. Methods used to solve such problem in cloud manufacturing are mainly based on the thought of transforming multiobjective optimization problem into single objective optimization problem, and thus use the mature single objective optimization algorithm to solve the converted problem among the existing research. However, many defects and deficiencies are existed in such transformation, which makes it difficult for users to obtain the global optimal resource service composition. So multiobjective optimization algorithms are used to study the problem of resource service composition optimal-selection directly, the main research contents are as follows:(1) Service encapsulation of cloud manufacturing resources. Firstly, the resources were classified based on the analysis of the characteristics of resources and the requirements of resource model in cloud manufacturing, and a formal description of equipment resource was carried out. Then the description information was mapped to a class template for virtualization, and a service encapsulation method of manufacturing equipment resources was presented. Finally, an example was used to verify the feasibility and convenience of this method.(2) Cloud manufacturing RSC modeling. Firstly, the problem of cloud manufacturing RSC optimal-selection problem was discussed in detail. Then the attribute indexes of resource service were analyzed, and the calculation method of attribute values of the RSC was given. Finally, a multiobjective mathematical model of the problem of RSC optimal-selection based on time, cost and satisfaction was constructed.(3) Method for solving the problem of RSC optimal-selection. Firstly, the defects and deficiencies of the traditional multiobjective optimization problem solving methods were analyzed,and multiobjective optimization algorithms were proposed to solve the problem of RSC optimal-selection. Then the implementation process of an improved fast elitist non-dominated sorting genetic algorithm was described in detail. Finally, an example was given to verify the feasibility and effectiveness of the proposed algorithm compared with some other algorithms.(4) Comprehensive evaluation of RSC in cloud manufacturing. Firstly, a six layer attribute index system evaluation model of RSC was established, and the method of preprocessing the attribute value was determined. Then a combinatorial weighting method based on the combination of analytic hierarchy process and entropy method was adopted to calculate weight coefficients. Finally, the combinatorial weighting method was applied to the improved TOPSIS and VIKOR method, and an example was used to verify the feasibility of these two kinds of multi-attribute decision making method in comprehensive evaluation of RSC in cloud manufacturing.