Abstract: The continuous casting process is the core and bottleneck process of steelmaking-continuous casting.Continuous casting machine maintenance is not only a necessary guarantee for the reliability of steelmaking-continuous casting production equipment,but also an important factor affecting the continuity of steelmaking-continuous casting production.The flexible maintenance requirement of continuous caster was introduced into the steelmaking-continuous casting scheduling.With the optimization goal of minimizing the waiting time between furnaces and the idle time of the equipment,a mixed integer programming model of the problem was established and a hybrid particle swarm optimization was proposed.Aiming at the shortcomings of particle swarm optimization that was easy to fall into the local optimum in the discrete problem,the cross and mutation strategy and the inverse learning strategy were designed.To improve the iteration efficiency,the simulation annealing search strategy of the global optimal particle was designed.Combined with the special production constraints and flexible maintenance constraints in the model,a reverse order solution rule was created.Comparative experiments based on actual production data had been carried out to prove the effectiveness of the model and algorithm.

Key words: steelmaking-continuous casting, production scheduling, flexible maintenance, particle swarm optimization, reverse learning

摘要： 连铸工序是炼钢连铸的核心,连铸机检修既是炼钢连铸生产设备可靠性的必要保障,也是影响炼钢连铸生产连续性的重要因素。鉴于此,将连铸机可检修时间为区间值的柔性检修要求引入炼钢连铸调度中,以最小化炉次间等待时间和设备闲置时间为优化目标,建立了问题的混合整数规划模型,并提出了一种混合粒子群算法。该算法针对粒子群算法在离散问题中容易陷入局部最优的缺点,设计了交叉变异策略和逆向学习策略;为提高迭代效率,设计了对全局最优粒子的模拟退火搜索策略;结合模型中的特殊生产约束和柔性检修约束特征,设计了一种逆序倒推的求解规则。基于实际生产数据进行了对比实验,证明了模型及算法的有效性。

关键词: 炼钢连铸, 生产调度, 柔性检修, 粒子群算法, 逆向学习