Research On Knowledge Meta Of Multi Disturbance Optimal Scheduling Knowledge Mining And Its System For Steel Production

Intelligent transformation and upgrading is an effective way for the sustainable development of steel industry.The application of new generation information technology characterized by ubiquitous technology in steel production scheduling makes the scheduling environment of steel enterprises present a ubiquitous manufacturing environment with rich information.Steel production scheduling in ubiquitous manufacturing environment has larger scale,more objectives and dynamic uncertainty,and the constraints are more complex and changeable.It is an NP-Hard problem that is not easy to solve.The traditional research on steel production scheduling can not meet the market demand of personalized,multi variety and small batch,and it is difficult to meet the intelligent demand of large-scale production and personalized customization of steel production in ubiquitous manufacturing environment.With the development of knowledge engineering,it is possible to study the steel production scheduling based on knowledge engineering and realize the integration,knowledge and intelligence of steel production scheduling system.Considering the multi disturbance events in the process of steel production scheduling,the multi disturbance scheduling problem of steel production based on knowledge engineering in ubiquitous manufacturing environment is studied.The development of knowledge-based intelligent scheduling system is of great significance to realize intelligent production scheduling in steel enterprises.Taking the multi disturbance events of steel production in the ubiquitous manufacturing environment as the object,the information flow and knowledge flow in the steel production process are analyzed and summarized by comprehensively using knowledge engineering,computer science theory and technology.The multi disturbance scheduling model of steel production in a ubiquitous manufacturing environment is established and the concept of knowledge meta is proposed.The component-based knowledge system of the multi disturbance scheduling of steel production is developed.The main research work is as follows.(1)Based on the ubiquitous manufacturing environment,combined with the information and automatic management and control system of steel enterprises,the rescheduling information flow of steel production are analyzed.With the information space theory,the multi disturbance rescheduling solution space of steel production in a ubiquitous manufacturing environment is described.It is divided into Time-dimensional,steel process information flow in the ubiquitous manufacturing environment,Devicedimensional,disturbance event perception and discovery in the ubiquitous manufacturing environment,and Information-dimensional,knowledge decision-making and optimization in the ubiquitous manufacturing environment.The ubiquitous information space model of multi disturbance scheduling in steel production is established.(2)The massive scheduling data and disturbance information in the ubiquitous manufacturing environment have the characteristics of multi-dimensional,multi granularity,and strong coupling.The concept of knowledge meta is proposed.The information objects of knowledge meta for steel production multi disturbance scheduling model are defined,including task knowledge,target knowledge,constraint knowledge,disturbance event knowledge,and resource knowledge.The knowledge meta of steel production multi disturbance scheduling model are represented by the ontology method.The ontology relation model of multi disturbance scheduling domain model knowledge of steel production is constructed,and the ontology of knowledge meta is instantiated.The reuse strategy of knowledge metais proposed,which includes knowledge description interface,knowledge behavior interface and knowledge release interface.A layer by layer encapsulation method of knowledge metais proposed,which abstracts the complex model,decomposes and encapsulates the knowledge models with different granularity.Taking the steelmaking continuous casting rescheduling model with start-up delay as an example,an application case of knowledge meta is given to verify the feasibility of knowledge meta representation.Based on component technology,the knowledge encapsulation process of steelmaking continuous casting rescheduling model under start-up delay is realized.(3)The knowledge decision table of multi disturbance events in steel production is constructed.The concepts of strong disturbance and weak disturbance are defined.The multi disturbance event characteristic index is selected from the furnace number,process,and equipment.Based on the C4.5 algorithm,the strong and weak disturbance classification method is established.A disturbance conflict resolution method based on the translation method under weak disturbance is proposed.Through the simulation case,the prediction accuracy of the proposed C4.5 algorithm is more than 80%,and the prediction average is 87%.Compared with the CART algorithm,it is verified that the proposed method is better.(4)According to the requirements of different disturbance scheduling models for different optimization algorithms,an improved multi-population quantum genetic algorithm(IMQGA)and a hyper-heuristic algorithm based on genetic programming(GPHH)are proposed to solve the multi disturbance scheduling model of steel production.IMQGA optimizes and searches multiple populations at the same time through multiple population improvement strategies.The improved quantum rotation angle strategy is used to give different rotation angles to different populations to update their quantum rotation gates,and give different evolutionary abilities to make them jump out of local optimization and achieve different search purposes.The production data of a steel plant is used for simulation verification.Compared with quantum genetic algorithm QGA and multipopulation quantum genetic algorithm MQGA,IMQGA has a faster convergence speed and better solution.At the same time,GP-HH is studied to generate an excellent heuristic scheduling rule set through genetic programming(GP).Through HH high-level strategy search and selection of heuristic rule set,the convergence speed of the algorithm is accelerated,and the global search ability of the algorithm is enhanced,which makes up for the deficiency that the rule-based heuristic algorithm can only obtain suboptimal solutions.Through the comparison of example simulation and heuristic rules,it is verified that the convergence speed and optimization ability of GP-HH are not as good as the heuristic scheduling method in small samples.Under larger samples,the performance of GP-HH is gradually better than the heuristic scheduling method,and a better solution can be obtained in an effective time.(5)According to the requirements of knowledge representation,knowledge mining of multi disturbance events and multi disturbance rescheduling of steel production multi disturbance scheduling model,a knowledge system prototype system of steel production multi disturbance scheduling is developed by using modeling tools such as Java,JSP,database,OWL and XML.Starting from the actual steel production process,the modules such as component packaging and disturbance rescheduling are realized,The knowledge description interface,knowledge release interface,and knowledge behavior interface are designed.The implementation process and method of model knowledge sub-module and model knowledge component management are described to verify the effectiveness of the component-based model knowledge space encapsulation method.Taking the production data of a steel plant as an example,the effectiveness of the multi disturbance event knowledge mining method and multi disturbance rescheduling algorithm is verified.