Research On Dynamic Task-aware Service Composition And Scheduling Optimization In Cloud Manufacturing

With the continuous development of economic globalization,an increasing number of manufacturers have been choosing to encapsulate manufacturing resources and capabilities into manufacturing services to enhance collaboration between enterprises to improve their competitiveness.Thus,the manufacturing industry has gradually transformed from the production-oriented mode to the service-oriented mode.Cloud manufacturing is a new service-oriented networked agile manufacturing mode which is efficient and knowledge-based.Cloud manufacturing technology can provide users with flexible and agile services through the unified and centralized intelligent management of various manufacturing resources.A cloud manufacturing service platform tends to receive multiple tasks concurrently,most of which are multi-granular and complex,and can be decomposed into several subtasks that need to be fulfilled by the cooperation among various services with different functionalities.Thus,a complex task should be decomposed into several subtasks which have their corresponding set of candidate manufacturing services with different quality of service values.While satisfying constraints of quality of service,services with various functionalities are aggregated and can be value-added.Appropriate services with different functionalities for achieving the corresponding tasks that optimize the service status and execution plans of manufacturing tasks are composed and scheduled.Thus,cloud manufacturing service composition and scheduling is a critical issue.However,a previous optimal execution plan may become non-optimal or even infeasible owing to the uncertainty of the real manufacturing environment where dynamic task is a vital source.This study takes the dynamic task as the research object,aiming to investigate manufacturing service composition and scheduling problem and achieve optimal manufacturing service allocation.At first,a single task-oriented cloud manufacturing dynamic composition model,and two service recomposition methods based on vertical collaboration and speed selection are proposed to deal with the urgent task request.Then,a multi-task-oriented cloud manufacturing dynamic scheduling model dealing with new task arrival is proposed,which considers both horizontal and vertical cooperation among service suppliers from supply chains.At last,according to the respective characters of the proposed models,based on the basicbiogeography-based optimization algorithm,an improved two-stage biogeography-based optimization algorithm and an improved multi-population biogeography-based optimization algorithm are proposed to improve the performance for solving the corresponding models.The main contributions of this paper can be concluded as follows:1.A single task-oriented cloud manufacturing dynamic composition model,and two service recomposition methods based on vertical collaboration and speed selection are proposed to expedite task completion.Further,an improved two-stage(i.e.,composition and recomposition)biogeography-based optimization algorithm is proposed to solve the corresponding model.A two-vector representation is proposed to make the biogeography-based optimization algorithm adaptive to the service recomposition problem.Then,the variable neighborhood search technique and elite replacement strategy are integrated to enhance the exploitation ability and convergence.Experiment results demonstrate that the proposed algorithm can obtain better cloud manufacturing service composition and recomposition solutions compared with the basic biogeography-based optimization algorithm,genetic algorithm,and differential evolution,and the two proposed service recomposition methods can reduce task execution time more effectively than the original composition method.2.A multi-task-oriented cloud manufacturing dynamic scheduling model that deals with new task arrival is proposed based on multi-supply chain mode,by considering both horizontal and vertical cooperation among service suppliers from supply chains.An improved multi-population biogeography-based optimization algorithm is proposed that uses a matrix-based solution representation and integrates the multi-population strategy,local search for the best solution,and the collaboration mechanism,for determining the optimal schedule.An illustrative case shows that a higher-quality schedule can be obtained when considering multi-supply chain collaboration than the conventional single-supply chain mode at both scheduling and rescheduling stages in the service scheduling model.Experiments are also conducted for verifying the effectiveness of the proposed algorithm for solving the presented model by comparing it with the basic biogeography-based optimization algorithm,genetic algorithm,and particle swarm optimization.