约束多目标粒子群算法在烧结配矿优化中的应用

Application of constrained multi-objective particle swarm optimization to sinter proportioning optimization

中文关键词  烧结配料  多目标粒子群算法  约束优化  区域划分

英文关键词  sintering proportioning process  multi-objective particle swarm optimization algorithm  constrained optimization  region division

中文摘要

      针对钢铁生产的烧结配料过程中铁矿石价格变化大、矿石品位波动、烧结原料信息复杂、烧结配料约束繁 多等对烧结配料成本的影响, 提出了一种基于区域划分的约束多目标粒子群优化算法对烧结配矿进行优化. 以国内 某钢铁厂的实际烧结生产配料为例, 构建烧结配料系统模型; 为了协调全局探索和局部搜索的关系, 将自适应角度 划分策略融入约束评价准则, 结合区域分布提取局部最优解信息, 同时引入双外部存储集机制, 维护种群多样性; 通 过标准函数集的测试, 验证了所提算法的有效性. 该算法应用于配矿过程中, 能够兼顾成本与全铁含量, 有效降低了 烧结配料的成本, 对烧结铁矿石资源的综合利用及质量保证具有重要意义.

英文摘要

      Aiming at the influence of iron ore price change, ore grade fluctuation, complex information of sintering raw materials and numerous constraints of sintering batching in iron and steel production, a constrained multi-objective particle swarm optimization algorithm based on region division is proposed to optimize the actual sintering production burden of a domestic iron and steel plant. The sintering burden system model is established. In order to coordinate the relationship between global development and local search, the adaptive angle division strategy is integrated into the constraint evaluation criterion, the local optimal solution information is extracted combined with regional distribution, and the two-archive mechanism is introduced to maintain population diversity. The effectiveness of the proposed algorithm is verified by the test of standard function set. The algorithm is applied to account the cost and total iron content in the process of ore blending, effectively reducing the cost of sintering ingredients. It has important significance for the comprehensive utilization of sintered iron ore resources and quality assurance.