基于改进人工蜂群算法的砂型铸造打磨工序并行机调度

Parallel Machine Scheduling of Foundry Grinding Process Based on Improved Artificial Bee Colony Algorithm

多品种小批量砂型铸造企业打磨工序任务调度是一种并行机调度问题,针对现有人工调度模式下任务分配不均衡、铸件在部分工人处积压、资源配置不合理等问题,提出了一种人工蜂群算法(Artificial Bee Colony algorithm,ABC)辅助的调度模式。首先,建立一种打磨工资系数计算方案,构建以最小化打磨工资系数和的差异、最小化每位工人打磨数量差异为优化目标的数学模型。然后,提出了改进的人工蜂群算法(Improved Discrete Artificial Bee Colony algorithm,IDABC),引入离散编码、采用启发式解码规则、使用防止近亲交配的交叉算子和嵌入禁忌搜索的侦察蜂等4个改进策略。再次,使用IDABC对5个不同规模的模拟数据集进行求解,实验结果显示IDABC在求解质量、求解速度和稳定性等方面均优于现有的标准ABC和遗传算法(Genetic Algorithm,GA)。最后,通过1个月内实际调度效果对比,表明了IDABC辅助调度模式能够有效提升打磨任务分配的均衡性,避免铸件在部分工人处堆积,实现资源的合理配置。

The task scheduling of grinding process in multi variety and small batch sand casting enterprises is a kind of parallel machine scheduling problem.Aiming at the problems of unbalanced task allocation,overstocking of castings at local workers and unreasonable resource allocation under the existing manual scheduling mode,an artificial bee colony algorithm(ABC)aided scheduling mode was proposed.Firstly,a calculation scheme of grinding wage coefficient is established,and a mathematical model is built to minimize the difference of grinding wage coefficient sum and grinding quantity difference of each worker.Then,an improved discrete artificial bee colony algorithm(IDABC)is proposed,which introduces discrete coding,adopts heuristic decoding rules,uses crossover operator to prevent inbreeding,and embeds a reconnaissance bee with tabu search.Thirdly,using IDABC to solve five different scale simulation data sets,the experimental results show that IDABC is superior to the existing standard ABC and genetic algorithm(GA)in solving quality,solving speed and stability.Finally,through the comparison of the actual scheduling effect in one month,it shows that IDABC auxiliary scheduling mode can effectively improve the balance of grinding task allocation,avoid casting accumulation in some workers,and realize the reasonable allocation of resources.