摘要
综合考虑联合补货与产品运输决策,针对在存储及装卸搬运等过程中引起的产品缺陷和上游供应商生产能力约束导致的部分产品缺货等实际生产中可能存在的问题,构建了存在产品缺陷和允许缺货的联合补货模型。模型同时考虑了资金、运输能力、以及"部分商品不能同时运输"这一运输条件等各方面的约束。由于该模型是一类混合整数非线性规划(MINLP)模型,为了提高求解效率,基于遗传算法与差分进化算法设计了混合算法,并与遗传算法进行了对比,证明了混合算法的寻优能力。对模型中的"成本"参数进行敏感度分析,结果表明库存持有成本、搬运费用、单位购买成本等参数的变动均会对模型结果造成影响,且单位购买成本对总成本的影响远大于库存持有成本和搬运费用。另外,构建了同等条件下不允许缺货的联合补货模型,对比优化结果发现在保证供应商生产能力的前提下,不缺货的联合补货总成本要低于缺货的补货成本。
The joint replenishment problem(JRP)is a multi-item inventory model of coordinating the replenishment of some items that can be ordered jointly from a single supplier.In such a system,the replenishment cost includes two components-a major ordering cost K incurred whenever an order is placed and a minor ordering cost ai incurred if item i is in the order.A JRP shares the major ordering cost,such as the fixed ordering cost and transportation cost,among the group of items that are ordered jointly.Hence,the cost can be reduced by 15%~30%compared to that of individually placing order policy.A joint replenishment model is presented to find the optimal order policy of a group of multiple items.This research involves defective items caused by imperfect production process or the damage occurred in transit.The supplier’s productivity is considered and the shortage is allowed.In addition,this research deals with allocating items in the trucks with budget constraint,transportation capacity constraint and shipment constraint.For example,prohibited items cannot be shipped together in the same truck.Since JRP is NP-hard problem,a hybrid genetic algorithm(HGA)is provided to determine an optimal replenishment cycle time T and the reorder frequencies mi for each item.Then numerical experiments are conducted to investigate how the model parameters,such as the percentage of defective items,demand rate,handling cost,holding cost and unit purchasing cost,affect the ordering policy.The results show that as the percentage of defective item increases,the multiplier of the order cycle T for each item,mi,remains the same and the expected total cost per unit time increases to compensate the defective items.The results also show that the total expected cost increases along with the demand rate,handling cost,holding cost and unit purchasing cost.Among them the demand rate and unit purchasing cost have a strong effect on the total cost.Therefore,accurate demand forecasting plays a very important role in practical supply chain management applications.Besides,the results reveal that the optimal family cycle length T will decrease when the demand rate and holding cost increases.In addition,optimal family cycle length will increase as the handling cost increases.The unit purchasing cost has little effect on the optimal family cycle length T.Furthermore,another JRP model where shortages are not allowed is developed and the results are compared to that of the JRP considering shortages.Results indicate that the total cost of the JRP,which is not out of stock,is obviously lower than that of JRP when the stock-out occurs.Given that stock-out is not allowed,the optimal replenishment cycle time T is shorter because order is placed more frequently to totally satisfy customers'demand.Finally,in order to test the performance of the proposed HGA,extensive computational experiments are conducted.HGA outperforms the classical genetic algorithm(GA)both in terms of the expected total cost and computational time.
作者
杨璐
陈彦如
杨洁
YANG Lu;CHEN Yan-ru;YANG Jie(School of Economics and Management,Southwest Jiaotong University,Chengdu 610031,China;School of Transportation and Logistics,Southwest Jiaotong University,Chengdu 610031,China)
出处
《管理工程学报》
CSSCI
CSCD
北大核心
2018年第4期195-203,共9页
Journal of Industrial Engineering and Engineering Management
基金
国家自然科学基金资助项目(71402149)
关键词
联合补货问题
缺货
缺陷率
生产能力
混合算法
Joint replenishment problem
Stock out
Defective item
Production capacity
Hybrid genetic algorithm
作者简介
杨璐(1992—),女,山西忻州人;西南交通大学经济管理学院硕士;研究方向:物流与供应链管理。
