Production Planning For Steelmaking-continuous Casting Based On Genetically Simulated Annealing Algorithm

Steelmaking is the most important process during production in iron and steelindustry. For steelmaking production organization, planning and scheduling is the corecontent. Planning and scheduling directly affect the efficiency of the production.Developing a feasible plan can balance equipment utilization and coordinate multistageproduction, which can improve the overall efficiency.The steelmaking planning is a typical combinatorial optimization NP-hard problem.As a research hotspot in recent years, remarkable progress was made both in theory andapplication. Many planning and scheduling systems were applied in enterprises withouteffectiveness. Therefore, it is necessary to conduct in-depth study of planningoptimization and system development about complex process.The method to solve planning problem has four steps：Problem analysis, key factorgain, modeling and solving, program implementation. Firstly, the article analysesproduction environment and requirement, and then abstracts problems characterized bymodel and strains. Through analyzing general steelmaking process, the article proposesproduction logistic network based on graph theory and builds a model by usingconfiguration modeling techniques. Secondly, depending on production network andactual product requirement, a mathematic model of production planning forsteelmaking-continues casting is established with optimization objective function whichminimizes flow time and waiting time. Thirdly, it is important to select appropriatealgorithm to solve the optimization model and ensure the efficiency and accuracy of thesolution results. The article chooses genetic simulation annealing algorithm to solve theproblem and gets feasible production plan. At last, based on model and algorithm, thearticle develops a planning simulation system to solve the actual problem. With UMLUnified Modeling tools and MVC framework model, the system encapsulated modelsand algorithms into corresponding functional module. The production planningsimulation system in article developed on Windows XP system Visual STDIO2005platform and testified models and algorithms in actual production conditions.The study shows that process network diagram can visually express dynamicrelationship and logical relationship between job tasks and work stations. OptimizationModel is use to ensure continuous casting constrain and minimize flow time and waittime, which made the model satisfies the requirement of practical production. Genetic algorithm was improved by matrix ending methods, which make it possible to solvemachine conflict problem during planning. Hybrid optimization algorithm can improvethe efficiency and feasibility of plan through combined with simulation annealingalgorithm that has great local search ability. UML and MVC framework was used toensure the expansion of the system and adaptability, which has a very important role insystem development and practical application.