Research In Steelmaking-Continuous Casting Scheduling Problem With Constraint Satisfaction And Genetic Algorithm

Steelmaking-continuous casting has complex production process, several production machines and crossed logistics path. In order to meet the requirements of steel enterprises such as multi-type and small batch, high quality and reasonable price, on time delivery, optimization model which is oriented to realistic environment and its algorithm must be established to realize reasonable process flow and quick process rhythm, so as to balance each process load, link logistics and production and shorten waiting time between procedures, then improve production efficiency and product quality, reduce the consumption of energy and materials.This paper defines steelmaking-continuous casting production scheduling problem as constraint satisfaction optimization problem, and establishes the constraint satisfaction optimization model. According to optimization model,main problem and subproblem based on Benders decomposition and its corresponding model are established to realize the simplified constraints. According to the idea of Benders decomposition method, process and code design of constraint satisfaction genetic algorithm (CSGA) are constructed using the constraint satisfaction technology in feasibility treatment and evolutionary mechanism of genetic.Based on the genetic evolution mechanism, This paper is designed mainly according to the population initialization, genetic structure, fitness function, decoding, termination conditions and parameter setting and other aspects. In feasibility processing based on constraint satisfaction, starting time intervals of charges are constructed. Main problems are solved by the solving rules and subproblems are solved through solution generating method of subproblems. Then establish conflict resolution technology based on limited buffer of rotating table, carry conflict resolution where the resource conflicts exist in subproblems from the conflict recognition and conflict resolution respectively and realize resolution of conflict with limited resources.According to the practical production data,Research shows that:the optimization model based on constraint satisfaction and the decomposition of the main problems and son problem based on Benders are reasonable、constraint satisfaction genetic algorithm and conflict resolution technology based on limited buffer of rotating table are effective. In the performance of production time combination for constraint satisfaction genetic algorithm, casting production time is the biggest impact factor, and with the opposite effect of no-casting production time. Comparing the present Solution with the makespan relaxation lower bound, constraint satisfaction genetic algorithm (CSGA) has certain nearly optimum.