Research On Flexible Job-shop Rescheduling Decision-making Method Based-on Steady-state Characteristic

With the increasing development of economic globalization, the production mode of modern manufacturing enterprises has shifted from traditional single-item big-lot pattern to multi-item small-lot pattern, gradually. In the meanwhile, the raising standards of market competition and the different preferences of customers’ consumption leads market demand to be more diverse, individual and customized. Consequently, enterprises have to pay more attention to the uncertainty existed in the production process and the influences it brings to the initial schedule. The uncertainties of abnormal events and information attributes entwine and couple, and resultingly, initial schedule deviates from actual production process, the efficiency and feasibility of initial schedule has been degraded. Therefore, rescheduling decision-making plays a big role in fast responding production disturbances, efficiently decreasing production cost, legitimately deploying manufacturing resources and optimally controlling production process.Aiming at the rescheduling decision-making problem in the flexible job-shop, firstly,the literature review of the related research field about rescheduling in production shop is surveyed. The uncertainty of production workshop has been analyzed, the production disturbances have been identified and the classification method is proposed, which enables the production disturbances classifying fastly. The meaning of rescheduling decision-making has been discussed here, and the definition of rescheduling steady-state is given, comprehensively investigating the rescheduling optimal method, driven mechanism and decision-making method in flexible job-shop while the relative problem models and the solution algorithms are put forward.The research contents and innovations of the dissertation are as follows:(1)On the basis of uncertain factors analysis of production workshop, the types of production disturbances have been identified. The characteristic index have been chosen from the dimension of jobs, operations and machines, and the classification method of production disturbances based on decision tree is proposed, which realizes the fast classification, and provide the research foundation for discussing rescheduling decision-making problem deeply.(2)The connotation and meaning of rescheduling decision-making has been discussed. The definition of rescheduling steady-state has been given. And the mathematic model is established, from the stable, efficient and economic perspective, respectively, in order to explore the changing rule of stability, efficiency and economic performance of production workshop after rescheduled through the mathematic description of rescheduling steady-state, which searches the reason that production system deviates from normal tendency and provides decision support for production management.(3)Aiming at the rescheduling optimization problem in flexible job-shop, the update rule of rescheduling problem model after rescheduled is given, and an hybrid estimation of distribution algorithm is put forward. In order to keep the diversification of population, k-means cluster is applied here to cluster the population. The individuals, chosen from each sub-cluster, composes the superior population, and the probability model that describes the distribution of solution space is established. Besides that, the variable neighborhood search is imported to the algorithm, avoiding trapping the local optimum.(4)Using rescheduling economic model to depict the phenomenon and changing rule of gain and loss during rescheduling in production workshop. Introducing cloud model to measure the uncertainty of rescheduling gain and loss. The improved reverse cloud algorithm is put forward to calculate the numeric characteristic of rescheduling gain and loss cloud which predicts the possible tendency of gain and loss after rescheduled at the current time. The rescheduling driven mechanism based on economic model is proposed, which make a judgment about whether or not reschedule through the possible tendency of gain and loss after rescheduled, and provide the evidence for rescheduling judgment.(5)The decision problem of optimal optional schedules in the flexible job-shop rescheduling is transformed in to a multi-attribute decision problem, which the attribute values real and the weight of attributes are linguistic variables. The 1-dimension normal cloud model is applied to describe the natural linguistic variables. Calculating the generalized distance between optional schedule in the set of reschedules and the positive(negative) ideal schedule based on several distance measurement, and using Borda count method realizes single-distance-sorting information fusion in order to acquire whole-distance-sorting information, which improves the expandability and flexibility of the decision model.