基于粒子群和约束满足的钢轧一体化批量计划优化

Integrated batch planning optimization based on particle swarm optimization and constraint satisfaction for steel production

在分析钢铁生产中的钢轧一体化批量计划编制问题基本特征的基础上给出了一体化编制策略,并建立了问题的约束满足优化模型。针对模型的NP难特性,提出了一种将改进离散粒子群算法、约束满足和邻域搜索相结合的混合算法。算法采用自然数矩阵编码,每个粒子代表其相应任务分配问题的解;在构造启发式解的基础上,利用邻域搜索方法计算粒子的适应值;为提高算法的收敛速度,利用约束满足技术生成初始化可行种群并修复迭代过程中产生的不可行解。基于企业实际生产数据的仿真实验结果验证了模型和算法的有效性。

The integrated batch planning of steel production continuous-casting hot-rolling was considered.By analyzing the essential operation properties of highlight problem,an integrated planning strategy was presented and a constraint satisfaction model was constructed.Considering the problem＇ s NP-hard feature,a hybrid algorithm combining Improved Discrete Particle Swarm Optimization（IDPSO）,constraint satisfaction and neighborhood search was proposed to solve this problem.With natural-number-matrix representation,each particle represented one solution to corresponding task allocation problem.Neighborhood search method was used to calculate particle＇ s fitness value based on heuristic solution of the sort optimization problem.To improve the algorithm＇ s convergence,the constraint satisfaction technique was employed to generate the initial feasible particle swarms and to revise unfeasible solutions during iterations.Validity of the model and algorithm were tested by calculating the data from production practices.