A Study On The Vertical Integration Of Supply Chain Optimization Models Based On Joint Replenishment Policy

Joint replenishment is an effective policy of procurement, which achieves cost saving through the sharing of the major ordering cost. It has more application value under the background of global procurement due to the higher distribution cost. As the key problem of inventory management, it has the closely link with the distribution strategy and the location strategy of distribution center, and influences each other. So, it has important theoretical and practical significance by considering joint replenishment strategy, delivery strategy, and location strategy integrated and designing the robustness and effectiveness algorithm.This thesis integrates the delivery and location problem of the supply chain to extend joint replenishment problem from the vertical. Combing the existing researches and the practice of the operation management, four integrated optimization model based on joint replenishment are proposed. In order to resolve these NP-hard, a hybrid self-adaptive differential evolution algorithm with wide generality and easy to implement is designed. The efficiency of the proposed algorithm is verified by the benchmark function testing. It provides the methodological support for the solving of the integrated models.Firstly, a joint replenishment and delivery model with grouping constraint is constructed based on the reality of the additional equipment cost occurs while the heterogeneous items transport together. The special case (penalty cost is zero) is designed to verified the model, then penalty cost is not equal to zero used to simulated the reality in experiment arrangement. Randomly generated ten problems from twelve combinations respectively verified the advantage in stableness and robustness of the proposed algorithm compare with genetic algorithm.Secondly, overcome the challenge of considering the vehicle routing simultaneously in the joint replenishment and delivery problem, a joint replenishment and delivery model with vehicle routing is proposed. Two scales problems verified the efficiency of the proposed algorithm. The contrastive analysis with the independent delivery showed that the independent delivery is well than joint delivery when the distance between the warehouse and the customers has little difference from the distance between the customers. In turn, the joint delivery is well.Thirdly, the joint replenishment and location-inventory model under constant demand reconstructed with more commonly. Three different scales of randomly numerical experiments were designed to analyze the performance of the proposed algorithm from the solution quality, computation time, and convergence speed. The results verified the efficiency and the advantage of the algorithm in the large scale problem. The sensitive analysis of cost parameters revealed that the location strategy has directly effect on replenishment policy, conversely the effect is little.Lastly, joint replenishment and location-inventory model under stochastic demand constructed due to the realistic of the demand of the customers is difficult to estimate. Based on the constant demand, a further analysis of the effect of the cost parameters on decisions is handled by enlarging the scope of varying. It gives the support to adjust cost parameters for enterprise. Furthermore, in order to analyze the role of joint replenishment policy, a contrastive analysis with independent replenishment policy was proposed. The results indicated that the joint replenishment policy always better than independent replenishment policy when the number of distribution is more than one. The enterprise can chose the reasonable replenishment policy according to the saving cost.