A Study On The Joint Replenishment Models Using Differential Evolution Algorithm

Inventory control is important for the normal operation of enterprise. Reasonable inventory management can enhance enterprise's competitiveness. Companies always have much variety of items to purchase from a single supplier or a same place when managing inventory. Joint replenishment strategy can reduce the annual ordering times, get more chance of price discount, and decrease the cost of transportation. This strategy is an effective way for cost savings. This thesis discussed several practical joint replenishment models and the novel algorithms for the models.Firstly, Joint Replenishment Problem (JRP) had been proved to be a typical NP-hard problem and the key is to find an algorithm that can solve this problem efficiently and effectively. Therefore, a differential evolution (DE) algorithm is presented to overcome the shortcomings of existing JRP methods. A modified algorithm is designed and the performance of improved algorithm is tested by four classic testing fuctions. Secondly, because the assumption of the determined JRP is unrealistic, two practical models, JRP with capital restriction and JRP under interdependence, are analyzised. The optimal replenishment strategy and the minimun total cost are given for the models. We also address the sensitivity analysis and investigate how misestimates of these parameters may influence the replenishment policy. At last, an extended joint replenishment and transportation model under fuzzy decision circumstance is constructed. The graded mean integration representation is used to defuzzify and DE algorithm is adopted to solve this model. A numerical example is given to verify the validity and superiority of this algorithm. Then we discuss the effect of fuzzy set on decision making and suggest that enterprises should exploit all the information and estimate the boundary of fuzzy set reasonably and scientifically.